问题描述
检查PostgreSQL表中存在哪些ID的速度如何
我写了以下函数,但当len(x)
大于500 000时,它会非常慢
import psycopg2
conn = psycopg2.connect(...)
cursor = conn.cursor(cursor_factory=psycopg2.extras.DictCursor)
def check_exist(lst):
exist = []
not_exist = []
for i in lst:
cursor.execute(f"SELECT * FROM table1 where id={i}")
row = cursor.fetchone()
if row:
exist.append(i)
else:
not_exist.append(i)
return exist, not_exist
x, y = check_exist(['2','4','3000','50000','10000000'])
推荐答案
您可以使用ANY,让postgres完成大部分工作:
import psycopg2, time
# database is in a datacenter just using a tunnel!
conn = psycopg2.connect("dbname=mf port=5959 host=localhost user=mf_usr")
cur = conn.cursor()
ids = [x for x in range(0, 750000)]
sql = """
SELECT array_agg(id) from __users where id = ANY(%s);
"""
# array_agg: Postgres returns an array of ids!
_start = time.time()
cur.execute(sql, (ids, ))
existingIds = cur.fetchone()[0]
missingIds = set(ids) - set(existingIds)
print(len(existingIds))
print(len(missingIds))
print('Took: %.6f seconds' % (time.time() - _start))
输出:
284365
465635
Took: 5.564851 seconds
注意:确保ID列上有索引