{{ header }}
Here we discuss a lot of the essential functionality common to the pandas data structures. To begin, let's create some example objects like we did in the :ref:`10 minutes to pandas <10min>` section:
.. ipython:: python
index = pd.date_range("1/1/2000", periods=8)
s = pd.Series(np.random.randn(5), index=["a", "b", "c", "d", "e"])
df = pd.DataFrame(np.random.randn(8, 3), index=index, columns=["A", "B", "C"])
To view a small sample of a Series or DataFrame object, use the :meth:`~DataFrame.head` and :meth:`~DataFrame.tail` methods. The default number of elements to display is five, but you may pass a custom number.
.. ipython:: python long_series = pd.Series(np.random.randn(1000)) long_series.head() long_series.tail(3)
pandas objects have a number of attributes enabling you to access the metadata
- shape: gives the axis dimensions of the object, consistent with ndarray
- Axis labels
- Series: index (only axis)
- DataFrame: index (rows) and columns
Note, these attributes can be safely assigned to!
.. ipython:: python df[:2] df.columns = [x.lower() for x in df.columns] df
pandas objects (:class:`Index`, :class:`Series`, :class:`DataFrame`) can be thought of as containers for arrays, which hold the actual data and do the actual computation. For many types, the underlying array is a :class:`numpy.ndarray`. However, pandas and 3rd party libraries may extend NumPy's type system to add support for custom arrays (see :ref:`basics.dtypes`).
To get the actual data inside a :class:`Index` or :class:`Series`, use
the .array property
.. ipython:: python s.array s.index.array
:attr:`~Series.array` will always be an :class:`~pandas.api.extensions.ExtensionArray`. The exact details of what an :class:`~pandas.api.extensions.ExtensionArray` is and why pandas uses them are a bit beyond the scope of this introduction. See :ref:`basics.dtypes` for more.
If you know you need a NumPy array, use :meth:`~Series.to_numpy` or :meth:`numpy.asarray`.
.. ipython:: python s.to_numpy() np.asarray(s)
When the Series or Index is backed by an :class:`~pandas.api.extensions.ExtensionArray`, :meth:`~Series.to_numpy` may involve copying data and coercing values. See :ref:`basics.dtypes` for more.
:meth:`~Series.to_numpy` gives some control over the dtype of the
resulting :class:`numpy.ndarray`. For example, consider datetimes with timezones.
NumPy doesn't have a dtype to represent timezone-aware datetimes, so there
are two possibly useful representations:
- An object-dtype :class:`numpy.ndarray` with :class:`Timestamp` objects, each
with the correct
tz - A
datetime64[ns]-dtype :class:`numpy.ndarray`, where the values have been converted to UTC and the timezone discarded
Timezones may be preserved with dtype=object
.. ipython:: python
ser = pd.Series(pd.date_range("2000", periods=2, tz="CET"))
ser.to_numpy(dtype=object)
Or thrown away with dtype='datetime64[ns]'
.. ipython:: python ser.to_numpy(dtype="datetime64[ns]")
Getting the "raw data" inside a :class:`DataFrame` is possibly a bit more
complex. When your DataFrame only has a single data type for all the
columns, :meth:`DataFrame.to_numpy` will return the underlying data:
.. ipython:: python df.to_numpy()
If a DataFrame contains homogeneously-typed data, the ndarray can actually be modified in-place, and the changes will be reflected in the data structure. For heterogeneous data (e.g. some of the DataFrame's columns are not all the same dtype), this will not be the case. The values attribute itself, unlike the axis labels, cannot be assigned to.
Note
When working with heterogeneous data, the dtype of the resulting ndarray will be chosen to accommodate all of the data involved. For example, if strings are involved, the result will be of object dtype. If there are only floats and integers, the resulting array will be of float dtype.
In the past, pandas recommended :attr:`Series.values` or :attr:`DataFrame.values`
for extracting the data from a Series or DataFrame. You'll still find references
to these in old code bases and online. Going forward, we recommend avoiding
.values and using .array or .to_numpy(). .values has the following
drawbacks:
- When your Series contains an :ref:`extension type <extending.extension-types>`, it's unclear whether :attr:`Series.values` returns a NumPy array or the extension array. :attr:`Series.array` will always return an :class:`~pandas.api.extensions.ExtensionArray`, and will never copy data. :meth:`Series.to_numpy` will always return a NumPy array, potentially at the cost of copying / coercing values.
- When your DataFrame contains a mixture of data types, :attr:`DataFrame.values` may involve copying data and coercing values to a common dtype, a relatively expensive operation. :meth:`DataFrame.to_numpy`, being a method, makes it clearer that the returned NumPy array may not be a view on the same data in the DataFrame.
pandas has support for accelerating certain types of binary numerical and boolean operations using
the numexpr library and the bottleneck libraries.
These libraries are especially useful when dealing with large data sets, and provide large
speedups. numexpr uses smart chunking, caching, and multiple cores. bottleneck is
a set of specialized cython routines that are especially fast when dealing with arrays that have
nans.
You are highly encouraged to install both libraries. See the section :ref:`Recommended Dependencies <install.recommended_dependencies>` for more installation info.
These are both enabled to be used by default, you can control this by setting the options:
pd.set_option("compute.use_bottleneck", False)
pd.set_option("compute.use_numexpr", False)With binary operations between pandas data structures, there are two key points of interest:
- Broadcasting behavior between higher- (e.g. DataFrame) and lower-dimensional (e.g. Series) objects.
- Missing data in computations.
We will demonstrate how to manage these issues independently, though they can be handled simultaneously.
DataFrame has the methods :meth:`~DataFrame.add`, :meth:`~DataFrame.sub`, :meth:`~DataFrame.mul`, :meth:`~DataFrame.div` and related functions :meth:`~DataFrame.radd`, :meth:`~DataFrame.rsub`, ... for carrying out binary operations. For broadcasting behavior, Series input is of primary interest. Using these functions, you can use to either match on the index or columns via the axis keyword:
.. ipython:: python
df = pd.DataFrame(
{
"one": pd.Series(np.random.randn(3), index=["a", "b", "c"]),
"two": pd.Series(np.random.randn(4), index=["a", "b", "c", "d"]),
"three": pd.Series(np.random.randn(3), index=["b", "c", "d"]),
}
)
df
row = df.iloc[1]
column = df["two"]
df.sub(row, axis="columns")
df.sub(row, axis=1)
df.sub(column, axis="index")
df.sub(column, axis=0)
Furthermore you can align a level of a MultiIndexed DataFrame with a Series.
.. ipython:: python
dfmi = df.copy()
dfmi.index = pd.MultiIndex.from_tuples(
[(1, "a"), (1, "b"), (1, "c"), (2, "a")], names=["first", "second"]
)
dfmi.sub(column, axis=0, level="second")
Series and Index also support the :func:`divmod` builtin. This function takes the floor division and modulo operation at the same time returning a two-tuple of the same type as the left hand side. For example:
.. ipython:: python s = pd.Series(np.arange(10)) s div, rem = divmod(s, 3) div rem idx = pd.Index(np.arange(10)) idx div, rem = divmod(idx, 3) div rem
We can also do elementwise :func:`divmod`:
.. ipython:: python div, rem = divmod(s, [2, 2, 3, 3, 4, 4, 5, 5, 6, 6]) div rem
In Series and DataFrame, the arithmetic functions have the option of inputting
a fill_value, namely a value to substitute when at most one of the values at
a location are missing. For example, when adding two DataFrame objects, you may
wish to treat NaN as 0 unless both DataFrames are missing that value, in which
case the result will be NaN (you can later replace NaN with some other value
using fillna if you wish).
.. ipython:: python df2 = df.copy() df2.loc["a", "three"] = 1.0 df df2 df + df2 df.add(df2, fill_value=0)
Series and DataFrame have the binary comparison methods eq, ne, lt, gt,
le, and ge whose behavior is analogous to the binary
arithmetic operations described above:
.. ipython:: python df.gt(df2) df2.ne(df)
These operations produce a pandas object of the same type as the left-hand-side
input that is of dtype bool. These boolean objects can be used in
indexing operations, see the section on :ref:`Boolean indexing<indexing.boolean>`.
You can apply the reductions: :attr:`~DataFrame.empty`, :meth:`~DataFrame.any`, :meth:`~DataFrame.all`.
.. ipython:: python (df > 0).all() (df > 0).any()
You can reduce to a final boolean value.
.. ipython:: python (df > 0).any().any()
You can test if a pandas object is empty, via the :attr:`~DataFrame.empty` property.
.. ipython:: python
df.empty
pd.DataFrame(columns=list("ABC")).empty
Warning
Asserting the truthiness of a pandas object will raise an error, as the testing of the emptiness or values is ambiguous.
.. ipython:: python
:okexcept:
if df:
print(True)
.. ipython:: python :okexcept: df and df2
See :ref:`gotchas<gotchas.truth>` for a more detailed discussion.
Often you may find that there is more than one way to compute the same
result. As a simple example, consider df + df and df * 2. To test
that these two computations produce the same result, given the tools
shown above, you might imagine using (df + df == df * 2).all(). But in
fact, this expression is False:
.. ipython:: python df + df == df * 2 (df + df == df * 2).all()
Notice that the boolean DataFrame df + df == df * 2 contains some False values!
This is because NaNs do not compare as equals:
.. ipython:: python np.nan == np.nan
So, NDFrames (such as Series and DataFrames) have an :meth:`~DataFrame.equals` method for testing equality, with NaNs in corresponding locations treated as equal.
.. ipython:: python (df + df).equals(df * 2)
Note that the Series or DataFrame index needs to be in the same order for equality to be True:
.. ipython:: python
df1 = pd.DataFrame({"col": ["foo", 0, np.nan]})
df2 = pd.DataFrame({"col": [np.nan, 0, "foo"]}, index=[2, 1, 0])
df1.equals(df2)
df1.equals(df2.sort_index())
You can conveniently perform element-wise comparisons when comparing a pandas data structure with a scalar value:
.. ipython:: python pd.Series(["foo", "bar", "baz"]) == "foo" pd.Index(["foo", "bar", "baz"]) == "foo"
pandas also handles element-wise comparisons between different array-like objects of the same length:
.. ipython:: python
pd.Series(["foo", "bar", "baz"]) == pd.Index(["foo", "bar", "qux"])
pd.Series(["foo", "bar", "baz"]) == np.array(["foo", "bar", "qux"])
Trying to compare Index or Series objects of different lengths will
raise a ValueError:
.. ipython:: python
:okexcept:
pd.Series(['foo', 'bar', 'baz']) == pd.Series(['foo', 'bar'])
pd.Series(['foo', 'bar', 'baz']) == pd.Series(['foo'])
A problem occasionally arising is the combination of two similar data sets where values in one are preferred over the other. An example would be two data series representing a particular economic indicator where one is considered to be of "higher quality". However, the lower quality series might extend further back in history or have more complete data coverage. As such, we would like to combine two DataFrame objects where missing values in one DataFrame are conditionally filled with like-labeled values from the other DataFrame. The function implementing this operation is :meth:`~DataFrame.combine_first`, which we illustrate:
.. ipython:: python
df1 = pd.DataFrame(
{"A": [1.0, np.nan, 3.0, 5.0, np.nan], "B": [np.nan, 2.0, 3.0, np.nan, 6.0]}
)
df2 = pd.DataFrame(
{
"A": [5.0, 2.0, 4.0, np.nan, 3.0, 7.0],
"B": [np.nan, np.nan, 3.0, 4.0, 6.0, 8.0],
}
)
df1
df2
df1.combine_first(df2)
The :meth:`~DataFrame.combine_first` method above calls the more general :meth:`DataFrame.combine`. This method takes another DataFrame and a combiner function, aligns the input DataFrame and then passes the combiner function pairs of Series (i.e., columns whose names are the same).
So, for instance, to reproduce :meth:`~DataFrame.combine_first` as above:
.. ipython:: python
def combiner(x, y):
return np.where(pd.isna(x), y, x)
df1.combine(df2, combiner)
There exists a large number of methods for computing descriptive statistics and other related operations on :ref:`Series <api.series.stats>`, :ref:`DataFrame <api.dataframe.stats>`. Most of these are aggregations (hence producing a lower-dimensional result) like :meth:`~DataFrame.sum`, :meth:`~DataFrame.mean`, and :meth:`~DataFrame.quantile`, but some of them, like :meth:`~DataFrame.cumsum` and :meth:`~DataFrame.cumprod`, produce an object of the same size. Generally speaking, these methods take an axis argument, just like ndarray.{sum, std, ...}, but the axis can be specified by name or integer:
- Series: no axis argument needed
- DataFrame: "index" (axis=0, default), "columns" (axis=1)
For example:
.. ipython:: python df df.mean(axis=0) df.mean(axis=1)
All such methods have a skipna option signaling whether to exclude missing
data (True by default):
.. ipython:: python df.sum(axis=0, skipna=False) df.sum(axis=1, skipna=True)
Combined with the broadcasting / arithmetic behavior, one can describe various statistical procedures, like standardization (rendering data zero mean and standard deviation of 1), very concisely:
.. ipython:: python ts_stand = (df - df.mean()) / df.std() ts_stand.std() xs_stand = df.sub(df.mean(axis=1), axis=0).div(df.std(axis=1), axis=0) xs_stand.std(axis=1)
Note that methods like :meth:`~DataFrame.cumsum` and :meth:`~DataFrame.cumprod`
preserve the location of NaN values. This is somewhat different from
:meth:`~DataFrame.expanding` and :meth:`~DataFrame.rolling` since NaN behavior
is furthermore dictated by a min_periods parameter.
.. ipython:: python df.cumsum()
Here is a quick reference summary table of common functions. Each also takes an
optional level parameter which applies only if the object has a
:ref:`hierarchical index<advanced.hierarchical>`.
| Function | Description |
|---|---|
count |
Number of non-NA observations |
sum |
Sum of values |
mean |
Mean of values |
median |
Arithmetic median of values |
min |
Minimum |
max |
Maximum |
mode |
Mode |
abs |
Absolute Value |
prod |
Product of values |
std |
Bessel-corrected sample standard deviation |
var |
Unbiased variance |
sem |
Standard error of the mean |
skew |
Sample skewness (3rd moment) |
kurt |
Sample kurtosis (4th moment) |
quantile |
Sample quantile (value at %) |
cumsum |
Cumulative sum |
cumprod |
Cumulative product |
cummax |
Cumulative maximum |
cummin |
Cumulative minimum |
Note that by chance some NumPy methods, like mean, std, and sum,
will exclude NAs on Series input by default:
.. ipython:: python np.mean(df["one"]) np.mean(df["one"].to_numpy())
:meth:`Series.nunique` will return the number of unique non-NA values in a Series:
.. ipython:: python series = pd.Series(np.random.randn(500)) series[20:500] = np.nan series[10:20] = 5 series.nunique()
There is a convenient :meth:`~DataFrame.describe` function which computes a variety of summary statistics about a Series or the columns of a DataFrame (excluding NAs of course):
.. ipython:: python
series = pd.Series(np.random.randn(1000))
series[::2] = np.nan
series.describe()
frame = pd.DataFrame(np.random.randn(1000, 5), columns=["a", "b", "c", "d", "e"])
frame.iloc[::2] = np.nan
frame.describe()
You can select specific percentiles to include in the output:
.. ipython:: python
series.describe(percentiles=[0.05, 0.25, 0.75, 0.95])
By default, the median is always included.
For a non-numerical Series object, :meth:`~Series.describe` will give a simple summary of the number of unique values and most frequently occurring values:
.. ipython:: python s = pd.Series(["a", "a", "b", "b", "a", "a", np.nan, "c", "d", "a"]) s.describe()
Note that on a mixed-type DataFrame object, :meth:`~DataFrame.describe` will restrict the summary to include only numerical columns or, if none are, only categorical columns:
.. ipython:: python
frame = pd.DataFrame({"a": ["Yes", "Yes", "No", "No"], "b": range(4)})
frame.describe()
This behavior can be controlled by providing a list of types as include/exclude
arguments. The special value all can also be used:
.. ipython:: python
frame.describe(include=["object"])
frame.describe(include=["number"])
frame.describe(include="all")
That feature relies on :ref:`select_dtypes <basics.selectdtypes>`. Refer to there for details about accepted inputs.
The :meth:`~DataFrame.idxmin` and :meth:`~DataFrame.idxmax` functions on Series and DataFrame compute the index labels with the minimum and maximum corresponding values:
.. ipython:: python s1 = pd.Series(np.random.randn(5)) s1 s1.idxmin(), s1.idxmax() df1 = pd.DataFrame(np.random.randn(5, 3), columns=["A", "B", "C"]) df1 df1.idxmin(axis=0) df1.idxmax(axis=1)
When there are multiple rows (or columns) matching the minimum or maximum value, :meth:`~DataFrame.idxmin` and :meth:`~DataFrame.idxmax` return the first matching index:
.. ipython:: python
df3 = pd.DataFrame([2, 1, 1, 3, np.nan], columns=["A"], index=list("edcba"))
df3
df3["A"].idxmin()
Note
idxmin and idxmax are called argmin and argmax in NumPy.
The :meth:`~Series.value_counts` Series method computes a histogram of a 1D array of values. It can also be used as a function on regular arrays:
.. ipython:: python data = np.random.randint(0, 7, size=50) data s = pd.Series(data) s.value_counts()
The :meth:`~DataFrame.value_counts` method can be used to count combinations across multiple columns.
By default all columns are used but a subset can be selected using the subset argument.
.. ipython:: python
data = {"a": [1, 2, 3, 4], "b": ["x", "x", "y", "y"]}
frame = pd.DataFrame(data)
frame.value_counts()
Similarly, you can get the most frequently occurring value(s), i.e. the mode, of the values in a Series or DataFrame:
.. ipython:: python
s5 = pd.Series([1, 1, 3, 3, 3, 5, 5, 7, 7, 7])
s5.mode()
df5 = pd.DataFrame(
{
"A": np.random.randint(0, 7, size=50),
"B": np.random.randint(-10, 15, size=50),
}
)
df5.mode()
Continuous values can be discretized using the :func:`cut` (bins based on values) and :func:`qcut` (bins based on sample quantiles) functions:
.. ipython:: python arr = np.random.randn(20) factor = pd.cut(arr, 4) factor factor = pd.cut(arr, [-5, -1, 0, 1, 5]) factor
:func:`qcut` computes sample quantiles. For example, we could slice up some normally distributed data into equal-size quartiles like so:
.. ipython:: python arr = np.random.randn(30) factor = pd.qcut(arr, [0, 0.25, 0.5, 0.75, 1]) factor
We can also pass infinite values to define the bins:
.. ipython:: python arr = np.random.randn(20) factor = pd.cut(arr, [-np.inf, 0, np.inf]) factor
To apply your own or another library's functions to pandas objects,
you should be aware of the three methods below. The appropriate
method to use depends on whether your function expects to operate
on an entire DataFrame or Series, row- or column-wise, or elementwise.
- Tablewise Function Application: :meth:`~DataFrame.pipe`
- Row or Column-wise Function Application: :meth:`~DataFrame.apply`
- Aggregation API: :meth:`~DataFrame.agg` and :meth:`~DataFrame.transform`
- Applying Elementwise Functions: :meth:`~DataFrame.map`
DataFrames and Series can be passed into functions.
However, if the function needs to be called in a chain, consider using the :meth:`~DataFrame.pipe` method.
First some setup:
.. ipython:: python
def extract_city_name(df):
"""
Chicago, IL -> Chicago for city_name column
"""
df["city_name"] = df["city_and_code"].str.split(",").str.get(0)
return df
def add_country_name(df, country_name=None):
"""
Chicago -> Chicago-US for city_name column
"""
col = "city_name"
df["city_and_country"] = df[col] + country_name
return df
df_p = pd.DataFrame({"city_and_code": ["Chicago, IL"]})
extract_city_name and add_country_name are functions taking and returning DataFrames.
Now compare the following:
.. ipython:: python
add_country_name(extract_city_name(df_p), country_name="US")
Is equivalent to:
.. ipython:: python
df_p.pipe(extract_city_name).pipe(add_country_name, country_name="US")
pandas encourages the second style, which is known as method chaining.
pipe makes it easy to use your own or another library's functions
in method chains, alongside pandas' methods.
In the example above, the functions extract_city_name and add_country_name each expected a DataFrame as the first positional argument.
What if the function you wish to apply takes its data as, say, the second argument?
In this case, provide pipe with a tuple of (callable, data_keyword).
.pipe will route the DataFrame to the argument specified in the tuple.
For example, we can fit a regression using statsmodels. Their API expects a formula first and a DataFrame as the second argument, data. We pass in the function, keyword pair (sm.ols, 'data') to pipe:
In [147]: import statsmodels.formula.api as sm
In [148]: bb = pd.read_csv("data/baseball.csv", index_col="id")
In [149]: (
.....: bb.query("h > 0")
.....: .assign(ln_h=lambda df: np.log(df.h))
.....: .pipe((sm.ols, "data"), "hr ~ ln_h + year + g + C(lg)")
.....: .fit()
.....: .summary()
.....: )
.....:
Out[149]:
<class 'statsmodels.iolib.summary.Summary'>
"""
OLS Regression Results
==============================================================================
Dep. Variable: hr R-squared: 0.685
Model: OLS Adj. R-squared: 0.665
Method: Least Squares F-statistic: 34.28
Date: Tue, 22 Nov 2022 Prob (F-statistic): 3.48e-15
Time: 05:34:17 Log-Likelihood: -205.92
No. Observations: 68 AIC: 421.8
Df Residuals: 63 BIC: 432.9
Df Model: 4
Covariance Type: nonrobust
===============================================================================
coef std err t P>|t| [0.025 0.975]
-------------------------------------------------------------------------------
Intercept -8484.7720 4664.146 -1.819 0.074 -1.78e+04 835.780
C(lg)[T.NL] -2.2736 1.325 -1.716 0.091 -4.922 0.375
ln_h -1.3542 0.875 -1.547 0.127 -3.103 0.395
year 4.2277 2.324 1.819 0.074 -0.417 8.872
g 0.1841 0.029 6.258 0.000 0.125 0.243
==============================================================================
Omnibus: 10.875 Durbin-Watson: 1.999
Prob(Omnibus): 0.004 Jarque-Bera (JB): 17.298
Skew: 0.537 Prob(JB): 0.000175
Kurtosis: 5.225 Cond. No. 1.49e+07
==============================================================================
Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 1.49e+07. This might indicate that there are
strong multicollinearity or other numerical problems.
"""
The pipe method is inspired by unix pipes and more recently dplyr and magrittr, which
have introduced the popular (%>%) (read pipe) operator for R.
The implementation of pipe here is quite clean and feels right at home in Python.
We encourage you to view the source code of :meth:`~DataFrame.pipe`.
Arbitrary functions can be applied along the axes of a DataFrame
using the :meth:`~DataFrame.apply` method, which, like the descriptive
statistics methods, takes an optional axis argument:
.. ipython:: python df.apply(lambda x: np.mean(x)) df.apply(lambda x: np.mean(x), axis=1) df.apply(lambda x: x.max() - x.min()) df.apply(np.cumsum) df.apply(np.exp)
The :meth:`~DataFrame.apply` method will also dispatch on a string method name.
.. ipython:: python
df.apply("mean")
df.apply("mean", axis=1)
The return type of the function passed to :meth:`~DataFrame.apply` affects the
type of the final output from DataFrame.apply for the default behaviour:
- If the applied function returns a
Series, the final output is aDataFrame. The columns match the index of theSeriesreturned by the applied function. - If the applied function returns any other type, the final output is a
Series.
This default behaviour can be overridden using the result_type, which
accepts three options: reduce, broadcast, and expand.
These will determine how list-likes return values expand (or not) to a DataFrame.
:meth:`~DataFrame.apply` combined with some cleverness can be used to answer many questions about a data set. For example, suppose we wanted to extract the date where the maximum value for each column occurred:
.. ipython:: python
tsdf = pd.DataFrame(
np.random.randn(1000, 3),
columns=["A", "B", "C"],
index=pd.date_range("1/1/2000", periods=1000),
)
tsdf.apply(lambda x: x.idxmax())
You may also pass additional arguments and keyword arguments to the :meth:`~DataFrame.apply` method.
.. ipython:: python
def subtract_and_divide(x, sub, divide=1):
return (x - sub) / divide
df_udf = pd.DataFrame(np.ones((2, 2)))
df_udf.apply(subtract_and_divide, args=(5,), divide=3)
Another useful feature is the ability to pass Series methods to carry out some Series operation on each column or row:
.. ipython:: python
tsdf = pd.DataFrame(
np.random.randn(10, 3),
columns=["A", "B", "C"],
index=pd.date_range("1/1/2000", periods=10),
)
tsdf.iloc[3:7] = np.nan
tsdf
tsdf.apply(pd.Series.interpolate)
Finally, :meth:`~DataFrame.apply` takes an argument raw which is False by default, which
converts each row or column into a Series before applying the function. When
set to True, the passed function will instead receive an ndarray object, which
has positive performance implications if you do not need the indexing
functionality.
The aggregation API allows one to express possibly multiple aggregation operations in a single concise way. This API is similar across pandas objects, see :ref:`groupby API <groupby.aggregate>`, the :ref:`window API <window.overview>`, and the :ref:`resample API <timeseries.aggregate>`. The entry point for aggregation is :meth:`DataFrame.aggregate`, or the alias :meth:`DataFrame.agg`.
We will use a similar starting frame from above:
.. ipython:: python
tsdf = pd.DataFrame(
np.random.randn(10, 3),
columns=["A", "B", "C"],
index=pd.date_range("1/1/2000", periods=10),
)
tsdf.iloc[3:7] = np.nan
tsdf
Using a single function is equivalent to :meth:`~DataFrame.apply`. You can also
pass named methods as strings. These will return a Series of the aggregated
output:
.. ipython:: python
tsdf.agg(lambda x: np.sum(x))
tsdf.agg("sum")
# these are equivalent to a ``.sum()`` because we are aggregating
# on a single function
tsdf.sum()
Single aggregations on a Series this will return a scalar value:
.. ipython:: python
tsdf["A"].agg("sum")
You can pass multiple aggregation arguments as a list.
The results of each of the passed functions will be a row in the resulting DataFrame.
These are naturally named from the aggregation function.
.. ipython:: python tsdf.agg(["sum"])
Multiple functions yield multiple rows:
.. ipython:: python tsdf.agg(["sum", "mean"])
On a Series, multiple functions return a Series, indexed by the function names:
.. ipython:: python tsdf["A"].agg(["sum", "mean"])
Passing a lambda function will yield a <lambda> named row:
.. ipython:: python tsdf["A"].agg(["sum", lambda x: x.mean()])
Passing a named function will yield that name for the row:
.. ipython:: python
def mymean(x):
return x.mean()
tsdf["A"].agg(["sum", mymean])
Passing a dictionary of column names to a scalar or a list of scalars, to DataFrame.agg
allows you to customize which functions are applied to which columns. Note that the results
are not in any particular order, you can use an OrderedDict instead to guarantee ordering.
.. ipython:: python
tsdf.agg({"A": "mean", "B": "sum"})
Passing a list-like will generate a DataFrame output. You will get a matrix-like output
of all of the aggregators. The output will consist of all unique functions. Those that are
not noted for a particular column will be NaN:
.. ipython:: python
tsdf.agg({"A": ["mean", "min"], "B": "sum"})
With .agg() it is possible to easily create a custom describe function, similar
to the built in :ref:`describe function <basics.describe>`.
.. ipython:: python from functools import partial q_25 = partial(pd.Series.quantile, q=0.25) q_25.__name__ = "25%" q_75 = partial(pd.Series.quantile, q=0.75) q_75.__name__ = "75%" tsdf.agg(["count", "mean", "std", "min", q_25, "median", q_75, "max"])
The :meth:`~DataFrame.transform` method returns an object that is indexed the same (same size)
as the original. This API allows you to provide multiple operations at the same
time rather than one-by-one. Its API is quite similar to the .agg API.
We create a frame similar to the one used in the above sections.
.. ipython:: python
tsdf = pd.DataFrame(
np.random.randn(10, 3),
columns=["A", "B", "C"],
index=pd.date_range("1/1/2000", periods=10),
)
tsdf.iloc[3:7] = np.nan
tsdf
Transform the entire frame. .transform() allows input functions as: a NumPy function, a string
function name or a user defined function.
.. ipython:: python
:okwarning:
tsdf.transform(np.abs)
tsdf.transform("abs")
tsdf.transform(lambda x: x.abs())
Here :meth:`~DataFrame.transform` received a single function; this is equivalent to a ufunc application.
.. ipython:: python np.abs(tsdf)
Passing a single function to .transform() with a Series will yield a single Series in return.
.. ipython:: python tsdf["A"].transform(np.abs)
Passing multiple functions will yield a column MultiIndexed DataFrame. The first level will be the original frame column names; the second level will be the names of the transforming functions.
.. ipython:: python tsdf.transform([np.abs, lambda x: x + 1])
Passing multiple functions to a Series will yield a DataFrame. The resulting column names will be the transforming functions.
.. ipython:: python tsdf["A"].transform([np.abs, lambda x: x + 1])
Passing a dict of functions will allow selective transforming per column.
.. ipython:: python
tsdf.transform({"A": np.abs, "B": lambda x: x + 1})
Passing a dict of lists will generate a MultiIndexed DataFrame with these selective transforms.
.. ipython:: python
:okwarning:
tsdf.transform({"A": np.abs, "B": [lambda x: x + 1, "sqrt"]})
Since not all functions can be vectorized (accept NumPy arrays and return another array or value), the methods :meth:`~DataFrame.map` on DataFrame and analogously :meth:`~Series.map` on Series accept any Python function taking a single value and returning a single value. For example:
.. ipython:: python
df4 = df.copy()
df4
def f(x):
return len(str(x))
df4["one"].map(f)
df4.map(f)
:meth:`Series.map` has an additional feature; it can be used to easily "link" or "map" values defined by a secondary series. This is closely related to :ref:`merging/joining functionality <merging>`:
.. ipython:: python
s = pd.Series(
["six", "seven", "six", "seven", "six"], index=["a", "b", "c", "d", "e"]
)
t = pd.Series({"six": 6.0, "seven": 7.0})
s
s.map(t)
:meth:`~Series.reindex` is the fundamental data alignment method in pandas. It is used to implement nearly all other features relying on label-alignment functionality. To reindex means to conform the data to match a given set of labels along a particular axis. This accomplishes several things:
- Reorders the existing data to match a new set of labels
- Inserts missing value (NA) markers in label locations where no data for that label existed
- If specified, fill data for missing labels using logic (highly relevant to working with time series data)
Here is a simple example:
.. ipython:: python s = pd.Series(np.random.randn(5), index=["a", "b", "c", "d", "e"]) s s.reindex(["e", "b", "f", "d"])
Here, the f label was not contained in the Series and hence appears as
NaN in the result.
With a DataFrame, you can simultaneously reindex the index and columns:
.. ipython:: python df df.reindex(index=["c", "f", "b"], columns=["three", "two", "one"])
Note that the Index objects containing the actual axis labels can be
shared between objects. So if we have a Series and a DataFrame, the
following can be done:
.. ipython:: python rs = s.reindex(df.index) rs rs.index is df.index
This means that the reindexed Series's index is the same Python object as the DataFrame's index.
:meth:`DataFrame.reindex` also supports an "axis-style" calling convention,
where you specify a single labels argument and the axis it applies to.
.. ipython:: python df.reindex(["c", "f", "b"], axis="index") df.reindex(["three", "two", "one"], axis="columns")
.. seealso:: :ref:`MultiIndex / Advanced Indexing <advanced>` is an even more concise way of doing reindexing.
Note
When writing performance-sensitive code, there is a good reason to spend
some time becoming a reindexing ninja: many operations are faster on
pre-aligned data. Adding two unaligned DataFrames internally triggers a
reindexing step. For exploratory analysis you will hardly notice the
difference (because reindex has been heavily optimized), but when CPU
cycles matter sprinkling a few explicit reindex calls here and there can
have an impact.
You may wish to take an object and reindex its axes to be labeled the same as another object. While the syntax for this is straightforward albeit verbose, it is a common enough operation that the :meth:`~DataFrame.reindex_like` method is available to make this simpler:
.. ipython:: python df2 = df.reindex(["a", "b", "c"], columns=["one", "two"]) df3 = df2 - df2.mean() df2 df3 df.reindex_like(df2)
The :meth:`~Series.align` method is the fastest way to simultaneously align two objects. It
supports a join argument (related to :ref:`joining and merging <merging>`):
join='outer': take the union of the indexes (default)join='left': use the calling object's indexjoin='right': use the passed object's indexjoin='inner': intersect the indexes
It returns a tuple with both of the reindexed Series:
.. ipython:: python s = pd.Series(np.random.randn(5), index=["a", "b", "c", "d", "e"]) s1 = s[:4] s2 = s[1:] s1.align(s2) s1.align(s2, join="inner") s1.align(s2, join="left")
For DataFrames, the join method will be applied to both the index and the columns by default:
.. ipython:: python df.align(df2, join="inner")
You can also pass an axis option to only align on the specified axis:
.. ipython:: python df.align(df2, join="inner", axis=0)
If you pass a Series to :meth:`DataFrame.align`, you can choose to align both
objects either on the DataFrame's index or columns using the axis argument:
.. ipython:: python df.align(df2.iloc[0], axis=1)
:meth:`~Series.reindex` takes an optional parameter method which is a
filling method chosen from the following table:
| Method | Action |
|---|---|
| ffill | Fill values forward |
| bfill | Fill values backward |
| nearest | Fill from the nearest index value |
We illustrate these fill methods on a simple Series:
.. ipython:: python
rng = pd.date_range("1/3/2000", periods=8)
ts = pd.Series(np.random.randn(8), index=rng)
ts2 = ts.iloc[[0, 3, 6]]
ts
ts2
ts2.reindex(ts.index)
ts2.reindex(ts.index, method="ffill")
ts2.reindex(ts.index, method="bfill")
ts2.reindex(ts.index, method="nearest")
These methods require that the indexes are ordered increasing or decreasing.
Note that the same result could have been achieved using
:ref:`ffill <missing_data.fillna>` (except for method='nearest') or
:ref:`interpolate <missing_data.interpolate>`:
.. ipython:: python ts2.reindex(ts.index).ffill()
:meth:`~Series.reindex` will raise a ValueError if the index is not monotonically increasing or decreasing. :meth:`~Series.fillna` and :meth:`~Series.interpolate` will not perform any checks on the order of the index.
The limit and tolerance arguments provide additional control over
filling while reindexing. Limit specifies the maximum count of consecutive
matches:
.. ipython:: python ts2.reindex(ts.index, method="ffill", limit=1)
In contrast, tolerance specifies the maximum distance between the index and indexer values:
.. ipython:: python ts2.reindex(ts.index, method="ffill", tolerance="1 day")
Notice that when used on a DatetimeIndex, TimedeltaIndex or
PeriodIndex, tolerance will coerced into a Timedelta if possible.
This allows you to specify tolerance with appropriate strings.
A method closely related to reindex is the :meth:`~DataFrame.drop` function.
It removes a set of labels from an axis:
.. ipython:: python df df.drop(["a", "d"], axis=0) df.drop(["one"], axis=1)
Note that the following also works, but is a bit less obvious / clean:
.. ipython:: python df.reindex(df.index.difference(["a", "d"]))
The :meth:`~DataFrame.rename` method allows you to relabel an axis based on some mapping (a dict or Series) or an arbitrary function.
.. ipython:: python s s.rename(str.upper)
If you pass a function, it must return a value when called with any of the labels (and must produce a set of unique values). A dict or Series can also be used:
.. ipython:: python
df.rename(
columns={"one": "foo", "two": "bar"},
index={"a": "apple", "b": "banana", "d": "durian"},
)
If the mapping doesn't include a column/index label, it isn't renamed. Note that extra labels in the mapping don't throw an error.
:meth:`DataFrame.rename` also supports an "axis-style" calling convention, where
you specify a single mapper and the axis to apply that mapping to.
.. ipython:: python
df.rename({"one": "foo", "two": "bar"}, axis="columns")
df.rename({"a": "apple", "b": "banana", "d": "durian"}, axis="index")
Finally, :meth:`~Series.rename` also accepts a scalar or list-like
for altering the Series.name attribute.
.. ipython:: python
s.rename("scalar-name")
The methods :meth:`DataFrame.rename_axis` and :meth:`Series.rename_axis`
allow specific names of a MultiIndex to be changed (as opposed to the
labels).
.. ipython:: python
df = pd.DataFrame(
{"x": [1, 2, 3, 4, 5, 6], "y": [10, 20, 30, 40, 50, 60]},
index=pd.MultiIndex.from_product(
[["a", "b", "c"], [1, 2]], names=["let", "num"]
),
)
df
df.rename_axis(index={"let": "abc"})
df.rename_axis(index=str.upper)
The behavior of basic iteration over pandas objects depends on the type. When iterating over a Series, it is regarded as array-like, and basic iteration produces the values. DataFrames follow the dict-like convention of iterating over the "keys" of the objects.
In short, basic iteration (for i in object) produces:
- Series: values
- DataFrame: column labels
Thus, for example, iterating over a DataFrame gives you the column names:
.. ipython:: python
df = pd.DataFrame(
{"col1": np.random.randn(3), "col2": np.random.randn(3)}, index=["a", "b", "c"]
)
for col in df:
print(col)
pandas objects also have the dict-like :meth:`~DataFrame.items` method to iterate over the (key, value) pairs.
To iterate over the rows of a DataFrame, you can use the following methods:
- :meth:`~DataFrame.iterrows`: Iterate over the rows of a DataFrame as (index, Series) pairs. This converts the rows to Series objects, which can change the dtypes and has some performance implications.
- :meth:`~DataFrame.itertuples`: Iterate over the rows of a DataFrame as namedtuples of the values. This is a lot faster than :meth:`~DataFrame.iterrows`, and is in most cases preferable to use to iterate over the values of a DataFrame.
Warning
Iterating through pandas objects is generally slow. In many cases, iterating manually over the rows is not needed and can be avoided with one of the following approaches:
- Look for a vectorized solution: many operations can be performed using built-in methods or NumPy functions, (boolean) indexing, ...
- When you have a function that cannot work on the full DataFrame/Series at once, it is better to use :meth:`~DataFrame.apply` instead of iterating over the values. See the docs on :ref:`function application <basics.apply>`.
- If you need to do iterative manipulations on the values but performance is important, consider writing the inner loop with cython or numba. See the :ref:`enhancing performance <enhancingperf>` section for some examples of this approach.
Warning
You should never modify something you are iterating over. This is not guaranteed to work in all cases. Depending on the data types, the iterator returns a copy and not a view, and writing to it will have no effect!
For example, in the following case setting the value has no effect:
.. ipython:: python
df = pd.DataFrame({"a": [1, 2, 3], "b": ["a", "b", "c"]})
for index, row in df.iterrows():
row["a"] = 10
df
Consistent with the dict-like interface, :meth:`~DataFrame.items` iterates through key-value pairs:
- Series: (index, scalar value) pairs
- DataFrame: (column, Series) pairs
For example:
.. ipython:: python
for label, ser in df.items():
print(label)
print(ser)
:meth:`~DataFrame.iterrows` allows you to iterate through the rows of a DataFrame as Series objects. It returns an iterator yielding each index value along with a Series containing the data in each row:
.. ipython:: python
for row_index, row in df.iterrows():
print(row_index, row, sep="\n")
Note
Because :meth:`~DataFrame.iterrows` returns a Series for each row, it does not preserve dtypes across the rows (dtypes are preserved across columns for DataFrames). For example,
.. ipython:: python df_orig = pd.DataFrame([[1, 1.5]], columns=["int", "float"]) df_orig.dtypes row = next(df_orig.iterrows())[1] row
All values in row, returned as a Series, are now upcasted
to floats, also the original integer value in column x:
.. ipython:: python row["int"].dtype df_orig["int"].dtype
To preserve dtypes while iterating over the rows, it is better to use :meth:`~DataFrame.itertuples` which returns namedtuples of the values and which is generally much faster than :meth:`~DataFrame.iterrows`.
For instance, a contrived way to transpose the DataFrame would be:
.. ipython:: python
df2 = pd.DataFrame({"x": [1, 2, 3], "y": [4, 5, 6]})
print(df2)
print(df2.T)
df2_t = pd.DataFrame({idx: values for idx, values in df2.iterrows()})
print(df2_t)
The :meth:`~DataFrame.itertuples` method will return an iterator yielding a namedtuple for each row in the DataFrame. The first element of the tuple will be the row's corresponding index value, while the remaining values are the row values.
For instance:
.. ipython:: python
for row in df.itertuples():
print(row)
This method does not convert the row to a Series object; it merely returns the values inside a namedtuple. Therefore, :meth:`~DataFrame.itertuples` preserves the data type of the values and is generally faster than :meth:`~DataFrame.iterrows`.
Note
The column names will be renamed to positional names if they are invalid Python identifiers, repeated, or start with an underscore. With a large number of columns (>255), regular tuples are returned.
Series has an accessor to succinctly return datetime like properties for the
values of the Series, if it is a datetime/period like Series.
This will return a Series, indexed like the existing Series.
.. ipython:: python
# datetime
s = pd.Series(pd.date_range("20130101 09:10:12", periods=4))
s
s.dt.hour
s.dt.second
s.dt.day
This enables nice expressions like this:
.. ipython:: python s[s.dt.day == 2]
You can easily produces tz aware transformations:
.. ipython:: python
stz = s.dt.tz_localize("US/Eastern")
stz
stz.dt.tz
You can also chain these types of operations:
.. ipython:: python
s.dt.tz_localize("UTC").dt.tz_convert("US/Eastern")
You can also format datetime values as strings with :meth:`Series.dt.strftime` which supports the same format as the standard :meth:`~datetime.datetime.strftime`.
.. ipython:: python
# DatetimeIndex
s = pd.Series(pd.date_range("20130101", periods=4))
s
s.dt.strftime("%Y/%m/%d")
.. ipython:: python
# PeriodIndex
s = pd.Series(pd.period_range("20130101", periods=4))
s
s.dt.strftime("%Y/%m/%d")
The .dt accessor works for period and timedelta dtypes.
.. ipython:: python
# period
s = pd.Series(pd.period_range("20130101", periods=4, freq="D"))
s
s.dt.year
s.dt.day
.. ipython:: python
# timedelta
s = pd.Series(pd.timedelta_range("1 day 00:00:05", periods=4, freq="s"))
s
s.dt.days
s.dt.seconds
s.dt.components
Note
Series.dt will raise a TypeError if you access with a non-datetime-like values.
Series is equipped with a set of string processing methods that make it easy to
operate on each element of the array. Perhaps most importantly, these methods
exclude missing/NA values automatically. These are accessed via the Series's
str attribute and generally have names matching the equivalent (scalar)
built-in string methods. For example:
.. ipython:: python
s = pd.Series(
["A", "B", "C", "Aaba", "Baca", np.nan, "CABA", "dog", "cat"], dtype="string"
)
s.str.lower()
Powerful pattern-matching methods are provided as well, but note that pattern-matching generally uses regular expressions by default (and in some cases always uses them).
Note
Prior to pandas 1.0, string methods were only available on object -dtype
Series. pandas 1.0 added the :class:`StringDtype` which is dedicated
to strings. See :ref:`text.types` for more.
Please see :ref:`Vectorized String Methods <text.string_methods>` for a complete description.
pandas supports three kinds of sorting: sorting by index labels, sorting by column values, and sorting by a combination of both.
The :meth:`Series.sort_index` and :meth:`DataFrame.sort_index` methods are used to sort a pandas object by its index levels.
.. ipython:: python
df = pd.DataFrame(
{
"one": pd.Series(np.random.randn(3), index=["a", "b", "c"]),
"two": pd.Series(np.random.randn(4), index=["a", "b", "c", "d"]),
"three": pd.Series(np.random.randn(3), index=["b", "c", "d"]),
}
)
unsorted_df = df.reindex(
index=["a", "d", "c", "b"], columns=["three", "two", "one"]
)
unsorted_df
# DataFrame
unsorted_df.sort_index()
unsorted_df.sort_index(ascending=False)
unsorted_df.sort_index(axis=1)
# Series
unsorted_df["three"].sort_index()
Sorting by index also supports a key parameter that takes a callable
function to apply to the index being sorted. For MultiIndex objects,
the key is applied per-level to the levels specified by level.
.. ipython:: python
s1 = pd.DataFrame({"a": ["B", "a", "C"], "b": [1, 2, 3], "c": [2, 3, 4]}).set_index(
list("ab")
)
s1
.. ipython:: python s1.sort_index(level="a") s1.sort_index(level="a", key=lambda idx: idx.str.lower())
For information on key sorting by value, see :ref:`value sorting <basics.sort_value_key>`.
The :meth:`Series.sort_values` method is used to sort a Series by its values. The
:meth:`DataFrame.sort_values` method is used to sort a DataFrame by its column or row values.
The optional by parameter to :meth:`DataFrame.sort_values` may used to specify one or more columns
to use to determine the sorted order.
.. ipython:: python
df1 = pd.DataFrame(
{"one": [2, 1, 1, 1], "two": [1, 3, 2, 4], "three": [5, 4, 3, 2]}
)
df1.sort_values(by="two")
The by parameter can take a list of column names, e.g.:
.. ipython:: python df1[["one", "two", "three"]].sort_values(by=["one", "two"])
These methods have special treatment of NA values via the na_position
argument:
.. ipython:: python s[2] = np.nan s.sort_values() s.sort_values(na_position="first")
Sorting also supports a key parameter that takes a callable function
to apply to the values being sorted.
.. ipython:: python s1 = pd.Series(["B", "a", "C"])
.. ipython:: python s1.sort_values() s1.sort_values(key=lambda x: x.str.lower())
key will be given the :class:`Series` of values and should return a Series
or array of the same shape with the transformed values. For DataFrame objects,
the key is applied per column, so the key should still expect a Series and return
a Series, e.g.
.. ipython:: python
df = pd.DataFrame({"a": ["B", "a", "C"], "b": [1, 2, 3]})
.. ipython:: python df.sort_values(by="a") df.sort_values(by="a", key=lambda col: col.str.lower())
The name or type of each column can be used to apply different functions to different columns.
Strings passed as the by parameter to :meth:`DataFrame.sort_values` may
refer to either columns or index level names.
.. ipython:: python
# Build MultiIndex
idx = pd.MultiIndex.from_tuples(
[("a", 1), ("a", 2), ("a", 2), ("b", 2), ("b", 1), ("b", 1)]
)
idx.names = ["first", "second"]
# Build DataFrame
df_multi = pd.DataFrame({"A": np.arange(6, 0, -1)}, index=idx)
df_multi
Sort by 'second' (index) and 'A' (column)
.. ipython:: python df_multi.sort_values(by=["second", "A"])
Note
If a string matches both a column name and an index level name then a warning is issued and the column takes precedence. This will result in an ambiguity error in a future version.
Series has the :meth:`~Series.searchsorted` method, which works similarly to :meth:`numpy.ndarray.searchsorted`.
.. ipython:: python ser = pd.Series([1, 2, 3]) ser.searchsorted([0, 3]) ser.searchsorted([0, 4]) ser.searchsorted([1, 3], side="right") ser.searchsorted([1, 3], side="left") ser = pd.Series([3, 1, 2]) ser.searchsorted([0, 3], sorter=np.argsort(ser))
Series has the :meth:`~Series.nsmallest` and :meth:`~Series.nlargest` methods which return the
smallest or largest n values. For a large Series this can be much
faster than sorting the entire Series and calling head(n) on the result.
.. ipython:: python s = pd.Series(np.random.permutation(10)) s s.sort_values() s.nsmallest(3) s.nlargest(3)
DataFrame also has the nlargest and nsmallest methods.
.. ipython:: python
df = pd.DataFrame(
{
"a": [-2, -1, 1, 10, 8, 11, -1],
"b": list("abdceff"),
"c": [1.0, 2.0, 4.0, 3.2, np.nan, 3.0, 4.0],
}
)
df.nlargest(3, "a")
df.nlargest(5, ["a", "c"])
df.nsmallest(3, "a")
df.nsmallest(5, ["a", "c"])
You must be explicit about sorting when the column is a MultiIndex, and fully specify
all levels to by.
.. ipython:: python
df1.columns = pd.MultiIndex.from_tuples(
[("a", "one"), ("a", "two"), ("b", "three")]
)
df1.sort_values(by=("a", "two"))
The :meth:`~DataFrame.copy` method on pandas objects copies the underlying data (though not the axis indexes, since they are immutable) and returns a new object. Note that it is seldom necessary to copy objects. For example, there are only a handful of ways to alter a DataFrame in-place:
- Inserting, deleting, or modifying a column.
- Assigning to the
indexorcolumnsattributes. - For homogeneous data, directly modifying the values via the
valuesattribute or advanced indexing.
To be clear, no pandas method has the side effect of modifying your data; almost every method returns a new object, leaving the original object untouched. If the data is modified, it is because you did so explicitly.
For the most part, pandas uses NumPy arrays and dtypes for Series or individual
columns of a DataFrame. NumPy provides support for float,
int, bool, timedelta64[ns] and datetime64[ns] (note that NumPy
does not support timezone-aware datetimes).
pandas and third-party libraries extend NumPy's type system in a few places. This section describes the extensions pandas has made internally. See :ref:`extending.extension-types` for how to write your own extension that works with pandas. See the ecosystem page for a list of third-party libraries that have implemented an extension.
The following table lists all of pandas extension types. For methods requiring dtype
arguments, strings can be specified as indicated. See the respective
documentation sections for more on each type.
pandas has two ways to store strings.
objectdtype, which can hold any Python object, including strings.- :class:`StringDtype`, which is dedicated to strings.
Generally, we recommend using :class:`StringDtype`. See :ref:`text.types` for more.
Finally, arbitrary objects may be stored using the object dtype, but should
be avoided to the extent possible (for performance and interoperability with
other libraries and methods. See :ref:`basics.object_conversion`).
A convenient :attr:`~DataFrame.dtypes` attribute for DataFrame returns a Series with the data type of each column.
.. ipython:: python
dft = pd.DataFrame(
{
"A": np.random.rand(3),
"B": 1,
"C": "foo",
"D": pd.Timestamp("20010102"),
"E": pd.Series([1.0] * 3).astype("float32"),
"F": False,
"G": pd.Series([1] * 3, dtype="int8"),
}
)
dft
dft.dtypes
On a Series object, use the :attr:`~Series.dtype` attribute.
.. ipython:: python dft["A"].dtype
If a pandas object contains data with multiple dtypes in a single column, the
dtype of the column will be chosen to accommodate all of the data types
(object is the most general).
.. ipython:: python # these ints are coerced to floats pd.Series([1, 2, 3, 4, 5, 6.0]) # string data forces an ``object`` dtype pd.Series([1, 2, 3, 6.0, "foo"])
The number of columns of each type in a DataFrame can be found by calling
DataFrame.dtypes.value_counts().
.. ipython:: python dft.dtypes.value_counts()
Numeric dtypes will propagate and can coexist in DataFrames.
If a dtype is passed (either directly via the dtype keyword, a passed ndarray,
or a passed Series), then it will be preserved in DataFrame operations. Furthermore,
different numeric dtypes will NOT be combined. The following example will give you a taste.
.. ipython:: python
df1 = pd.DataFrame(np.random.randn(8, 1), columns=["A"], dtype="float64")
df1
df1.dtypes
df2 = pd.DataFrame(
{
"A": pd.Series(np.random.randn(8), dtype="float32"),
"B": pd.Series(np.random.randn(8)),
"C": pd.Series(np.random.randint(0, 255, size=8), dtype="uint8"), # [0,255] (range of uint8)
}
)
df2
df2.dtypes
By default integer types are int64 and float types are float64,
regardless of platform (32-bit or 64-bit).
The following will all result in int64 dtypes.
.. ipython:: python
pd.DataFrame([1, 2], columns=["a"]).dtypes
pd.DataFrame({"a": [1, 2]}).dtypes
pd.DataFrame({"a": 1}, index=list(range(2))).dtypes
Note that Numpy will choose platform-dependent types when creating arrays.
The following WILL result in int32 on 32-bit platform.
.. ipython:: python frame = pd.DataFrame(np.array([1, 2]))
Types can potentially be upcasted when combined with other types, meaning they are promoted
from the current type (e.g. int to float).
.. ipython:: python df3 = df1.reindex_like(df2).fillna(value=0.0) + df2 df3 df3.dtypes
:meth:`DataFrame.to_numpy` will return the lower-common-denominator of the dtypes, meaning the dtype that can accommodate ALL of the types in the resulting homogeneous dtyped NumPy array. This can force some upcasting.
.. ipython:: python df3.to_numpy().dtype
You can use the :meth:`~DataFrame.astype` method to explicitly convert dtypes from one to another. These will by default return a copy,
even if the dtype was unchanged (pass copy=False to change this behavior). In addition, they will raise an
exception if the astype operation is invalid.
Upcasting is always according to the NumPy rules. If two different dtypes are involved in an operation, then the more general one will be used as the result of the operation.
.. ipython:: python
df3
df3.dtypes
# conversion of dtypes
df3.astype("float32").dtypes
Convert a subset of columns to a specified type using :meth:`~DataFrame.astype`.
.. ipython:: python
dft = pd.DataFrame({"a": [1, 2, 3], "b": [4, 5, 6], "c": [7, 8, 9]})
dft[["a", "b"]] = dft[["a", "b"]].astype(np.uint8)
dft
dft.dtypes
Convert certain columns to a specific dtype by passing a dict to :meth:`~DataFrame.astype`.
.. ipython:: python
dft1 = pd.DataFrame({"a": [1, 0, 1], "b": [4, 5, 6], "c": [7, 8, 9]})
dft1 = dft1.astype({"a": np.bool_, "c": np.float64})
dft1
dft1.dtypes
Note
When trying to convert a subset of columns to a specified type using :meth:`~DataFrame.astype` and :meth:`~DataFrame.loc`, upcasting occurs.
:meth:`~DataFrame.loc` tries to fit in what we are assigning to the current dtypes, while [] will overwrite them taking the dtype from the right hand side. Therefore the following piece of code produces the unintended result.
.. ipython:: python
dft = pd.DataFrame({"a": [1, 2, 3], "b": [4, 5, 6], "c": [7, 8, 9]})
dft.loc[:, ["a", "b"]].astype(np.uint8).dtypes
dft.loc[:, ["a", "b"]] = dft.loc[:, ["a", "b"]].astype(np.uint8)
dft.dtypes
pandas offers various functions to try to force conversion of types from the object dtype to other types.
In cases where the data is already of the correct type, but stored in an object array, the
:meth:`DataFrame.infer_objects` and :meth:`Series.infer_objects` methods can be used to soft convert
to the correct type.
.. ipython:: python
import datetime
df = pd.DataFrame(
[
[1, 2],
["a", "b"],
[datetime.datetime(2016, 3, 2), datetime.datetime(2016, 3, 2)],
]
)
df = df.T
df
df.dtypes
Because the data was transposed the original inference stored all columns as object, which
infer_objects will correct.
.. ipython:: python df.infer_objects().dtypes
The following functions are available for one dimensional object arrays or scalars to perform hard conversion of objects to a specified type:
:meth:`~pandas.to_numeric` (conversion to numeric dtypes)
.. ipython:: python m = ["1.1", 2, 3] pd.to_numeric(m)
:meth:`~pandas.to_datetime` (conversion to datetime objects)
.. ipython:: python import datetime m = ["2016-07-09", datetime.datetime(2016, 3, 2)] pd.to_datetime(m)
:meth:`~pandas.to_timedelta` (conversion to timedelta objects)
.. ipython:: python m = ["5us", pd.Timedelta("1day")] pd.to_timedelta(m)
To force a conversion, we can pass in an errors argument, which specifies how pandas should deal with elements
that cannot be converted to desired dtype or object. By default, errors='raise', meaning that any errors encountered
will be raised during the conversion process. However, if errors='coerce', these errors will be ignored and pandas
will convert problematic elements to pd.NaT (for datetime and timedelta) or np.nan (for numeric). This might be
useful if you are reading in data which is mostly of the desired dtype (e.g. numeric, datetime), but occasionally has
non-conforming elements intermixed that you want to represent as missing:
.. ipython:: python
:okwarning:
import datetime
m = ["apple", datetime.datetime(2016, 3, 2)]
pd.to_datetime(m, errors="coerce")
m = ["apple", 2, 3]
pd.to_numeric(m, errors="coerce")
m = ["apple", pd.Timedelta("1day")]
pd.to_timedelta(m, errors="coerce")
In addition to object conversion, :meth:`~pandas.to_numeric` provides another argument downcast, which gives the
option of downcasting the newly (or already) numeric data to a smaller dtype, which can conserve memory:
.. ipython:: python
m = ["1", 2, 3]
pd.to_numeric(m, downcast="integer") # smallest signed int dtype
pd.to_numeric(m, downcast="signed") # same as 'integer'
pd.to_numeric(m, downcast="unsigned") # smallest unsigned int dtype
pd.to_numeric(m, downcast="float") # smallest float dtype
As these methods apply only to one-dimensional arrays, lists or scalars; they cannot be used directly on multi-dimensional objects such as DataFrames. However, with :meth:`~pandas.DataFrame.apply`, we can "apply" the function over each column efficiently:
.. ipython:: python
import datetime
df = pd.DataFrame([["2016-07-09", datetime.datetime(2016, 3, 2)]] * 2, dtype="O")
df
df.apply(pd.to_datetime)
df = pd.DataFrame([["1.1", 2, 3]] * 2, dtype="O")
df
df.apply(pd.to_numeric)
df = pd.DataFrame([["5us", pd.Timedelta("1day")]] * 2, dtype="O")
df
df.apply(pd.to_timedelta)
Performing selection operations on integer type data can easily upcast the data to floating.
The dtype of the input data will be preserved in cases where nans are not introduced.
See also :ref:`Support for integer NA <gotchas.intna>`.
.. ipython:: python
dfi = df3.astype("int32")
dfi["E"] = 1
dfi
dfi.dtypes
casted = dfi[dfi > 0]
casted
casted.dtypes
While float dtypes are unchanged.
.. ipython:: python
dfa = df3.copy()
dfa["A"] = dfa["A"].astype("float32")
dfa.dtypes
casted = dfa[df2 > 0]
casted
casted.dtypes
The :meth:`~DataFrame.select_dtypes` method implements subsetting of columns
based on their dtype.
First, let's create a :class:`DataFrame` with a slew of different dtypes:
.. ipython:: python
df = pd.DataFrame(
{
"string": list("abc"),
"int64": list(range(1, 4)),
"uint8": np.arange(3, 6).astype("u1"),
"float64": np.arange(4.0, 7.0),
"bool1": [True, False, True],
"bool2": [False, True, False],
"dates": pd.date_range("now", periods=3),
"category": pd.Series(list("ABC")).astype("category"),
}
)
df["tdeltas"] = df.dates.diff()
df["uint64"] = np.arange(3, 6).astype("u8")
df["other_dates"] = pd.date_range("20130101", periods=3)
df["tz_aware_dates"] = pd.date_range("20130101", periods=3, tz="US/Eastern")
df
And the dtypes:
.. ipython:: python df.dtypes
:meth:`~DataFrame.select_dtypes` has two parameters include and exclude that allow you to
say "give me the columns with these dtypes" (include) and/or "give the
columns without these dtypes" (exclude).
For example, to select bool columns:
.. ipython:: python df.select_dtypes(include=[bool])
You can also pass the name of a dtype in the NumPy dtype hierarchy:
.. ipython:: python df.select_dtypes(include=["bool"])
:meth:`~pandas.DataFrame.select_dtypes` also works with generic dtypes as well.
For example, to select all numeric and boolean columns while excluding unsigned integers:
.. ipython:: python df.select_dtypes(include=["number", "bool"], exclude=["unsignedinteger"])
To select string columns you must use the object dtype:
.. ipython:: python df.select_dtypes(include=["object"])
To see all the child dtypes of a generic dtype like numpy.number you
can define a function that returns a tree of child dtypes:
.. ipython:: python
def subdtypes(dtype):
subs = dtype.__subclasses__()
if not subs:
return dtype
return [dtype, [subdtypes(dt) for dt in subs]]
All NumPy dtypes are subclasses of numpy.generic:
.. ipython:: python
subdtypes(np.generic)
Note
pandas also defines the types category, and datetime64[ns, tz], which are not integrated into the normal
NumPy hierarchy and won't show up with the above function.