格式CREATE TABLE查询和标识符列表

时间:2019-02-15 17:47:53

标签: python python-3.x postgresql psycopg2

我想用Python编写一个从CSV创建PostgreSQL表的脚本。与其使用psycopg2.copy_from,而是希望我提供更个性化,更灵活的功能。

很显然,我将读取CSV文件的第一行并从中获取列名称的列表。然后我想将此列表转换为在postgreSQL中创建表时要编写的列列表:

"column_name_1" text,
"column_name_2" text,
"column_name_3" text,
...
"column_name_N" text

(默认情况下,我希望我的所有列都具有文本类型;稍后可能会更改这些类型)。 请注意,我确实希望在列名周围出现双引号,因为这些列名可能包含空格或带重音的字符。

我尝试使用psycopg2.sql.Identifier在列名两边加上双引号,但是失败了,因为我想将这些标识符与“文本”连接起来。

这是我到目前为止尝试过的:

import psycopg2
import csv 

conn = psycopg2.connect(
                            dbname = "spatialDB",host="localhost",
                            port = 5432, user = "postgres",
                            password="postgres"
                            )

cur = conn.cursor()

FileName = "csv_to_import.csv"
file = open(FileName,'r',encoding='utf-8')
reader = csv.reader(file,delimiter=";")

columns = next(reader)

# The line below is what I wanted to do, but you cannot concatenate 
# an SQL identifier with a string 

#column_types = [psycopg2.sql.Identifier(c) + " text" for c in colums]

# This is what I did instead but it's ugly
column_types = ["\"" + c + "\"  text" for c in columns]

schema = "myschema"
table_name = "import_csv_python"

model_query = "CREATE TABLE {0}.{1}({2});" 
column_list = ",".join(column_types)

query = model_query.format(schema,table_name,column_list)


cur.execute(query)
file.close()
cur.close()
conn.commit()

您如何纠正此代码以使其使用psycopg2.sql.Identifier函数正确引用列名? (对于参数{0}和{1},分别是模式和表名,也有同样的问题)

1 个答案:

答案 0 :(得分:1)

使用方法as_string(context)构建列列表,并将所有字符串参数转换为Composable:

import psycopg2.sql as sql

column_types = [sql.Identifier(c).as_string(cur) + " text" for c in columns]

schema = sql.Identifier("myschema")
table_name = sql.Identifier("import_csv_python")

model_query = sql.SQL("CREATE TABLE {0}.{1}({2});") 
column_list = sql.SQL(",".join(column_types))

query = model_query.format(schema,table_name,column_list)