从 sqlite3 确定最大列数 [英] Determine maximum number of columns from sqlite3
问题描述
是否有可能在运行时获得 sqlite3 支持的最大列数?此数据库限制是通过编译时变量 SQLITE_MAX_COLUMN 建立的(请参阅 limits一>).默认值通常为 2000 列.
Is it possible to get the maximum number of columns supported from sqlite3 at runtime? This database limitation is established with a compile-time variable SQLITE_MAX_COLUMN (see limits). The default is normally 2000 columns.
我正在寻找可从 Python 或 SQL 界面访问的内容.
I'm looking for something accessible from either the Python or SQL interface.
推荐答案
这在实践中似乎是不可能的(即，如果没有按照 dan04 相当出色的答案采用非常昂贵的蛮力方法).
This appears to be impossible in practical terms (i.e. without a very expensive brute-force approach along the lines of dan04's rather brilliant answer).
删除>
来源(1、2) 用于 sqlite3 模块不包含对 SQLITE_MAX_COLUMN 或一般编译时间限制；似乎也没有任何方法可以从 SQL 界面中访问它们.
UPDATE:
更新:
对dan04 的解决方案进行简单修改以使用二分搜索可大大加快速度:
import sqlite3
def max_columns():
db = sqlite3.connect(':memory:')
low = 1
high = 32767 # hard limit <http://www.sqlite.org/limits.html>
while low < high - 1:
guess = (low + high) // 2
try:
db.execute('CREATE TABLE T%d (%s)' % (
guess, ','.join('C%d' % i for i in range(guess))
))
except sqlite3.DatabaseError as ex:
if 'too many columns' in str(ex):
high = guess
else:
raise
else:
low = guess
return low
Running the code above through timeit.repeat():通过timeit.repeat()运行上面的代码:
... 平均运行时间为 30.76/150 = 0.205 秒(在 2.6 GHz 四核机器上) - 不完全快，但可能比 15 更有用-20 秒的踢它直到它打破"从一个方法开始计数.
... which comes to an average run-time of 30.76 / 150 = 0.205 seconds (on a 2.6 GHz quad-core machine) - not exactly fast, but likely more usable than the 15-20 seconds of the "kick it 'til it breaks" counting-from-one method.
这篇关于从 sqlite3 确定最大列数的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
