Unable to call iterdump method for Sqlite3 connections

[kesry]

Bug report

Bug description:

import sqlite3


def dict_factory(cursor, row):
    d = {}
    for idx, col in enumerate(cursor.description):
        d[col[0]] = row[idx]
    return d


def main():
    conn = sqlite3.connect(":memory:")
    cur = conn.cursor()
    # conn.row_factory = sqlite3.Row # is ok
    conn.row_factory = dict_factory # is fail
    cur.executescript("""    
        create table if not exists test(
            id integer primary key AUTOINCREMENT
        );
    """)
    cur.close()
    for line in conn.iterdump():
        print(line)


main()

CPython versions tested on:

3.12

Operating systems tested on:

Windows

Linked PRs

• gh-118221: Always use the default row factory in sqlite3.iterdump() #118223
• [3.12] gh-118221: Always use the default row factory in sqlite3.iterdump() (#118223) #118270

[erlend-aasland]

FTR, the traceback is:

Traceback (most recent call last):
  File "../iterdump.py", line 26, in <module>
    main()
  File "../iterdump.py", line 22, in main
    for line in conn.iterdump():
  File "/opt/homebrew/Cellar/[email protected]/3.8.19/Frameworks/Python.framework/Versions/3.8/lib/python3.8/sqlite3/dump.py", line 55, in _iterdump
    query_res = cu.execute(q)
sqlite3.OperationalError: no such table: name

Happens on Python versions from 3.8 through 3.13. Using the sqlite3.Row factory seems to work fine.

[kesry]

FTR, the traceback is:

Traceback (most recent call last):
  File "../iterdump.py", line 26, in <module>
    main()
  File "../iterdump.py", line 22, in main
    for line in conn.iterdump():
  File "/opt/homebrew/Cellar/[email protected]/3.8.19/Frameworks/Python.framework/Versions/3.8/lib/python3.8/sqlite3/dump.py", line 55, in _iterdump
    query_res = cu.execute(q)
sqlite3.OperationalError: no such table: name

Happens on Python versions from 3.8 through 3.13. Using the sqlite3.Row factory seems to work fine.

sqlite3. Row can't meet the needs of the production environment, is there any other way?

[erlend-aasland]

The problem is that iterdump does not work with a custom row factory like this. Perhaps we should save the row factory before doing iterdump and restore it upon completion.

[kesry]

iterdump

Well, Thanks!

[erlend-aasland]

I still think we should fix this. Perhaps a mention in the docs is sufficient.

[erlend-aasland]

The problem is that iterdump does not work with a custom row factory like this. Perhaps we should save the row factory before doing iterdump and restore it upon completion.

The solution is easier: overriding the row factory of a cursor does not impact the row factory of the parent connection, so all we need to do is to make sure the cursor used by iterdump() uses the default row factory by explicitly setting cursor.row_factory = None.

[erlend-aasland]

Resolved by:

• gh-118221: Always use the default row factory in sqlite3.iterdump() #118223
• [3.12] gh-118221: Always use the default row factory in sqlite3.iterdump() (#118223) #118270 (auto-merge enabled)

Thanks for the report, @kesry!

[kesry]

Resolved by:

• gh-118221: Always use the default row factory in sqlite3.iterdump() #118223
• [3.12] gh-118221: Always use the default row factory in sqlite3.iterdump() (#118223) #118270 (auto-merge enabled)

Thanks for the report, @kesry!

Greet!👍