如何在 python 的特定场景中忽略 csv 分隔符？

Question

我正在尝试使用 CSV 文件在 DB 中插入数据。

import psycopg2 #import the postgres library
#connect to the database
conn = psycopg2.connect(host='1.11.11.111',
                   dbname='postgres',
                   user='postgres',
                   password='myPassword',
                   port='1234')  
#create a cursor object 
#cursor object is used to interact with the database
cur = conn.cursor()
#open the csv file using python standard file I/O
#copy file into the table just created 
with open("C:/Users/Harshal/Desktop/tar.csv", 'r') as f:
next(f) 
cur.copy_from(f, 'geotargets_india',sep=',')
conn.commit()
conn.close()
f.close()

我的表如下：

create table public.geotargets_india(
Criteria_ID integer not null,
Name character varying(50) COLLATE pg_catalog."default" NOT NULL,
Canonical_Name character varying(100) COLLATE pg_catalog."default" NOT NULL,
Parent_ID NUMERIC(10,2),
Country_Code character varying(10) COLLATE pg_catalog."default" NOT NULL,
Target_Type character varying(50) COLLATE pg_catalog."default" NOT NULL,
Status character varying(50) COLLATE pg_catalog."default" NOT NULL
)

我的 CSV 看起来像：

我得到的错误是： 例如，如果仔细查看我的 csv 行：1007740,Hyderabad,"Hyderabad,Telangana,India",9061642.0,IN,City,Active 。在这里，Canonical_Name 有“，”分隔的字符串，这会导致错误，并假设 CSV 中的列比表中的多。如何解决这个问题？ 注意：我假设错误只是由于这个原因。 CSV Link

Answer 1

foo.csv：

It is header which will be ignored------------------------------------
1007740,Hyderabad,"Hyderabad,Telangana,India",9061642.0,IN,City,Active

Python：

import psycopg2
conn = psycopg2.connect('')
cur = conn.cursor()
f = open('foo.csv', 'r')
cur.copy_expert("""copy geotargets_india from stdin with (format csv, header, delimiter ',', quote '"')""", f)
conn.commit()

psql：

table geotargets_india;
┌─────────────┬───────────┬───────────────────────────┬────────────┬──────────────┬─────────────┬────────┐
│ criteria_id │   name    │      canonical_name       │ parent_id  │ country_code │ target_type │ status │
├─────────────┼───────────┼───────────────────────────┼────────────┼──────────────┼─────────────┼────────┤
│     1007740 │ Hyderabad │ Hyderabad,Telangana,India │ 9061642.00 │ IN           │ City        │ Active │
└─────────────┴───────────┴───────────────────────────┴────────────┴──────────────┴─────────────┴────────┘

Answer 2

Canonical_Name 中的问题是正确的。我成功地将行1007740,Hyderabad,"Hyderabad",9061642.0,IN,City,Active 导入到您的结构表中。

很遗憾，copy_from 方法不支持 csv 分隔符参数。这是文档https://www.psycopg.org/docs/cursor.html#cursor.copy_from

因此您可以使用制表符分隔符重新格式化 csv 文件，然后使用 copy_from

import csv
import psycopg2 #import the postgres library
#connect to the database
conn = psycopg2.connect(host='1.11.11.111',
                   dbname='postgres',
                   user='postgres',
                   password='myPassword',
                   port='1234')
#create a cursor object
#cursor object is used to interact with the database
cur = conn.cursor()
#open the csv file using python standard file I/O
#copy file into the table just created

with open("C:/Users/Harshal/Desktop/tar.csv", 'r') as f:
    reader = csv.reader(f, delimiter=",")
    with open("C:/Users/Harshal/Desktop/tar.tsv", 'w') as tsv:
        writer = csv.writer(tsv, delimiter='\t')
        writer.writerows(reader)

with open("C:/Users/Harshal/Desktop/tar.tsv", 'r') as f:
    next(f)
    cur.copy_from(f, 'geotargets_india',sep='\t')
    conn.commit()
    conn.close()
    f.close()

Answer 3

您应该自己在 Python 中读取和解析 CSV 文件，然后使用 INSERT 语句将数据加载到数据库中。

import csv
import psycopg2

conn = psycopg2.connect(
    host='1.11.11.111',
    dbname='postgres',
    user='postgres',
    password='myPassword',
    port='1234'
)  
cur = conn.cursor()

with open("tar.csv") as fd:
    rdr = csv.DictReader(fd)
    cur.executemany("""
        INSERT INTO geotargets_india
        VALUES (%(Criteria_ID)s, %(Name)s, %(Canonical_Name)s, %(Parent_ID)s, %(Country_Code)s, %(Target_Type)s, %(Status)s);
        """,
        rdr
    )

cur.close()
conn.close()

上面有几个cmets。 csv.DictReader 类将返回您的 CSV 字典。返回的 DictReader 对象 rdr 是可迭代的，因此可以直接在 psycopg2 的 cursor.executemany 函数中使用，这可能比自己遍历 csv DictReader 对象效率更高。