psycopg2 sql.SQL: 组合元素必须是可组合的

Question

import psycopg2
from psycopg2 import sql
import datetime


def connect(conn):
    """ Connect to the PostgreSQL database server """
    # create a cursor
    cur = conn.cursor()

    # https://www.psycopg.org/docs/sql.html

    col_names = sql.SQL(', ').join([
        sql.Identifier('lookup'),
        sql.Identifier('LOL1'),
        sql.Identifier('LOL2'),
        sql.Identifier('identifier1'),
        sql.Identifier('identifier2'),
        sql.Identifier('timestamp_first_entry'),
        sql.Identifier('timestamp_last_entry'),
        sql.Identifier('some_number')
    ]),

    values = sql.SQL(', ').join([
        "hash",
        [  # LOL1
            [[1.0, 2.0],  # list of lists of lists 1.1
             [3.0, 0.5]],
            [[-1.0, -2.0], # list of lists of lists 1.2
             [-3.0, -4.0]]
        ], [  # LOL2
            [[1.0, 2.0],  # list of lists of lists 2.1
             [3.0, 0.5]],
            [[-1.0, -2.0], # list of lists of lists 2.2
             [-3.0, -4.0]]
        ],
        "identifier1",
        "identifier2",
        datetime.datetime(2021, 5, 10),
        datetime.datetime(2021, 5, 12),
        20
    ])

    query_base = sql.SQL("insert into {table_name}({col_names} values ({values}))".format(
        table_name=sql.Identifier("raw_orderbook"),
        col_names=col_names,
        values=values
    ))

    cur.execute(query_base)

    cur.close()


# https://www.psycopg.org/docs/sql.html

if __name__ == '__main__':
    conn = psycopg2.connect(
        host="localhost",
        database="test",
        user="ian",
        password="")

    connect(conn)

导致错误

Traceback (most recent call last):
  File "/home/ian/PycharmProjects/panthera/src/database/tester.py", line 73, in <module>
    connect(conn)
  File "/home/ian/PycharmProjects/panthera/src/database/tester.py", line 33, in connect
    values = sql.SQL(', ').join([
  File "/home/ian/anaconda3/envs/panthera/lib/python3.9/site-packages/psycopg2/sql.py", line 288, in join
    return Composed(rv)
  File "/home/ian/anaconda3/envs/panthera/lib/python3.9/site-packages/psycopg2/sql.py", line 109, in __init__
    raise TypeError(
TypeError: Composed elements must be Composable, got 'lookup' instead

鉴于 composable 是基础 class，我不太确定错误的原因。据我所知，根据 docs 中的示例，似乎是正确的。我的 sql_schema

DROP TABLE IF EXISTS "raw_orderbooks";
CREATE TABLE "raw_orderbooks"
(

    lookup                text, -- essentially our lookup key comprising: hash(exchange_pair_timestapOfFirstEntry)
    LOL1                  DOUBLE PRECISION[][][],
    LOL2                  DOUBLE PRECISION[][][],
    identifier1                  text,
    identifier2            text,
    timestamp_first_entry TIMESTAMP,
    timestamp_last_entry  TIMESTAMP,
    some_number       int
);

我的另一个想法是将非日期时间值包装在 sql.Literal 中，但后来我运行进入日期时间问题。

环顾四周，我看到其他人使用不同的符号，但感觉不那么“干净”，这绝对是主观的，但 IDK。

编辑

根据回答的评论，我更新的脚本是

import psycopg2
from psycopg2 import sql
import datetime


def connect(conn):
    """ Connect to the PostgreSQL database server """
    # create a cursor
    cur = conn.cursor()

    # https://www.psycopg.org/docs/sql.html

    col_names = sql.SQL(', ').join([
        sql.Identifier('lookup'),
        sql.Identifier('LOL1'),
        sql.Identifier('LOL2'),
        sql.Identifier('identifier1'),
        sql.Identifier('identifier2'),
        sql.Identifier('timestamp_first_entry'),
        sql.Identifier('timestamp_last_entry'),
        sql.Identifier('some_number')
    ]),

    values = sql.SQL(', ').join([
        "hash",
        [  # LOL1
            [[1.0, 2.0],  # list of lists of lists 1.1
             [3.0, 0.5]],
            [[-1.0, -2.0], # list of lists of lists 1.2
             [-3.0, -4.0]]
        ], [  # LOL2
            [[1.0, 2.0],  # list of lists of lists 2.1
             [3.0, 0.5]],
            [[-1.0, -2.0], # list of lists of lists 2.2
             [-3.0, -4.0]]
        ],
        "identifier1",
        "identifier2",
        datetime.datetime(2021, 5, 10),
        datetime.datetime(2021, 5, 12),
        20
    ])
    table_col_names = ['lookup','LOL1','LOL2','identifier1','identifier2','timestamp_first_entry','timestamp_last_entry','some_number']

    col_names = sql.SQL(', ').join(sql.Identifier(n) for n in table_col_names )

    place_holders = sql.SQL(', ').join(sql.Placeholder() * len(table_col_names ))

    query_base = sql.SQL("insert into {table_name} ({col_names}) values ({values})").format(
        table_name=sql.Identifier("raw_orderbook"),
        col_names=col_names,
        values=place_holders
    )

    print(query_base.as_string(conn))
    cur.execute(query_base,values)

    cur.close(

但我仍在运行关注这个问题TypeError: Composed elements must be Composable, got 'lookup' instead

Answer 1

sql.SQL.join 加入 Composable 的序列，但您在 values 变量中提供普通变量，如 string 、 list 等。

其次，我建议使用 placeholder 作为查询中的值，然后传递 values 来执行。

table_col_names = ['lookup','LOL1','LOL2','identifier1','identifier2','timestamp_first_entry','timestamp_last_entry','some_number']

col_names = sql.SQL(', ').join(sql.Identifier(n) for n in table_col_names )

place_holders = sql.SQL(', ').join(sql.Placeholder() * len(table_col_names ))

query_base = sql.SQL("insert into {table_name} ({col_names}) values ({values})").format(
        table_name=sql.Identifier("raw_orderbook"),
        col_names=col_names,
        values=place_holders
    )

print(query_base.as_string(conn))
cur.execute(query_base,values)

psycopg2 sql.SQL: 组合元素必须是可组合的

psycopg2 sql.SQL: Composed elements must be Composable

python

sql

postgresql

psycopg2

编辑