关于Python内置sqlite3模块依赖机制及Ubuntu环境下链接状态检测的技术问询

Great question! Let's break this down, using Python 3.5 on Ubuntu as an example (the logic applies to newer Python versions too).

First, let's clarify the default behavior: Ubuntu's official Python packages (like the 3.5 version you mentioned) dynamically link the system-wide libsqlite3.so shared library for the _sqlite3 extension module. But custom-compiled Python builds might statically link SQLite instead—so here are reliable ways to verify which one your setup uses:

Method 1: Check shared library dependencies with ldd #

The most straightforward way is to inspect the _sqlite3 shared object directly using the ldd command, which lists dynamic dependencies.

Run this in your terminal:

ldd /usr/lib/python3.5/lib-dynload/_sqlite3.cpython-35m-x86_64-linux-gnu.so

• If you see a line like libsqlite3.so.0 => /usr/lib/x86_64-linux-gnu/libsqlite3.so.0 (0x00007f...), that means the module is dynamically linking the system's SQLite shared library.
• If there's no mention of libsqlite3.so in the output, the module was built with SQLite statically linked.

Method 2: Compare SQLite versions (Python vs. system) #

You can cross-check the SQLite version used by Python against the system's installed version:

1. In a Python shell, run:

   import sqlite3
   print("Python's SQLite version:", sqlite3.sqlite_version)
   

2. Check the system's SQLite version in the terminal:

   sqlite3 --version  # Or dpkg -l libsqlite3-0 to get the package version
   

• If the versions match, it's almost certainly using the system's shared library.
• If they differ, the Python module was likely built with a statically linked SQLite copy (common in custom Python builds).

Method 3: Inspect dynamic symbols with objdump #

Another way to confirm dynamic dependencies is using objdump to check the "NEEDED" flags of the shared object:

objdump -x /usr/lib/python3.5/lib-dynload/_sqlite3.cpython-35m-x86_64-linux-gnu.so | grep NEEDED

• A line reading NEEDED libsqlite3.so.0 confirms dynamic linking.
• No such line means static linking.

Quick recap of the import chain #

As you noted, when you import sqlite3 in Python 3.5, the call flow is:

import sqlite3 → sqlite3/__init__.py: from sqlite3.dbapi2 import * → sqlite3/dbapi2.py: from _sqlite3 import * → loads the shared object /usr/lib/python3.5/lib-dynload/_sqlite3.cpython-35m-x86_64-linux-gnu.so

This shared object is the actual bridge between Python and SQLite—so checking its dependencies tells you everything you need to know.

内容的提问来源于stack exchange，提问作者Basj