Question

我一直在尝试使用Python中的SQL进行一些CSV文件合并，SQL查询如下：

WITH
MATCHES AS( -- get all matches
    SELECT      CSV2.*
                , CSV1.ROW as ROW_1                 
                , CSV1.C4 as C4_1
                , CSV1.C5 as C5_1
    FROM        CSV2 
    LEFT JOIN   CSV1 
    ON          CSV1.C4 LIKE '%' || CSV2.C2 || '%'    
),
EXACT AS( -- matches where CSV1.C4 = CSV1.C5
    SELECT      *
    FROM        MATCHES
    WHERE       C4_1 = C5_1
),
MIN_ROW AS( -- CSV1.ROW of first occurence for each CSV2.C1
    SELECT      C1
                , min(ROW_1) as ROW_1
    FROM        MATCHES
    WHERE       C1 NOT IN (SELECT C1 FROM EXACT)
    GROUP BY    C1, C2, C3, C4, C5                  
)
-- use C4=C5 first
SELECT      *
FROM        EXACT
UNION
-- if match not in exact, use first occurence
SELECT      MATCHES.*
FROM        MIN_ROW
INNER JOIN  MATCHES
ON          MIN_ROW.C1 = MATCHES.C1
AND         (MIN_ROW.ROW_1 = MATCHES.ROW_1 OR MIN_ROW.ROW_1 IS NULL)
ORDER BY    C1

但是，我一直收到OperationalError：near＆＃34; WITH＆＃34;：语法错误。我对SQL并不十分熟悉，并且之前没有尝试过在Python中使用它。可以做些什么来解决这个错误？ Update1-第一对括号内的查询工作正常，没有＆＃34;匹配为＆＃34;部分。以下是示例csv文件： CSV1

data13      data23      d      main_data1;main_data2      data13         data23
data12      data22      d      main_data1;main_data2      data12         data22
data11      data21      d      main_data1;main_data2      data11         data21
data3       data4       d      main_data2;main_data4      data3          data4
data52      data62      d      main_data3                 data51         data62
data51      data61      d      main_data3                 main_data3     data61
data7       data8       d      main_data4                 data7          data8

CSV2

id1      main_data1      a1      a2      a3
id2      main_data2      b1      b2      b3
id3      main_data3      c1      c2      c3
id4      main_data4      d1      d2      d3
id5      main_data5      e1      e2      e3

和Python代码

import csv
import sqlite3

def createTable(cursor, rows, tablename):
    tableCreated = False
    for row in rows:
        if not tableCreated:
            sql = "CREATE TABLE %s(ROW INTEGER PRIMARY KEY, " + ", ".join(["c%d" % (i+1) for i in range(len(row))]) + ")"
            cur.execute(sql % tablename)
            tableCreated = True
        sql = "INSERT INTO %s VALUES(NULL, " + ", ".join(["'" + c + "'" for c in row]) + ")"
        cur.execute(sql % tablename)
    conn.commit()


conn = sqlite3.connect(":memory:")
cur = conn.cursor()

for filename, tablename in [(path_to_csv1, "CSV1"), (path_to_csv2, "CSV2")]:
    with open(filename, "r") as f:
        reader = csv.reader(f, delimiter=',')        
        rows = [row for row in reader]
    createTable(cur, rows, tablename)

Update2-我使用Python 2.7

Answer 1

自SQLite 3.8.3起支持WITH子句。 Python 2.7附带的版本要老得多。

更新到最新的Python，或使用其他数据库驱动程序，如apsw，或者，如果使用动态库编译Python，请尝试替换SQLite库文件。

Answer 2

添加';'在With之前

;WITH
MATCHES AS( ...

Python中的SQL查询给出OperationalError：“WITH”附近：语法错误

2 个答案: