我一直在追逐这个错误一周:
pyodbc.Error: ('HY000', '[HY000] [Vertica][VerticaDSII] (20) An error occurred during query execution: (20) (SQLFetch)')
经过大量的反复试验,关键是此特定SELECT
查询(简单的问题)的组合,加上curs.execute()
后commit()
的调用。我真的很想知道您的想法。下面(接下来是配置信息)就是一个示例。了解为什么发生这种情况的任何帮助将不胜感激。我很乐意提供更多细节。
版本:
- python:3.3.2 [GCC 4.4.6 20120305(Red Hat 4.4.6-4)]在Linux上
- unixODBC:2.2.14
- Vertica:Vertica Analytic数据库V7.0.0-0
- OS:Linux版本2.6.32-279.14.1.el6.x86_64(mockbuild@c6b8.bsys.dev.centos.org)(GCC版本4.4.6 20120305(Red Hat 4.4.6-4)(GCC)(GCC))
来源:
import pyodbc
query = '''
SELECT relVarTable0.id AS id, relVarTable1.val, relVarTable2.val, relVarTable3.val, relVarTable4.val
FROM (SELECT id FROM B) relVarTable0
LEFT JOIN
(SELECT B1.id AS id, AVG(A1.X) AS val
FROM B B1, AB AB1, A A1
WHERE B1.id = AB1.B_id AND A1.id = AB1.A_id AND A1.X IS NOT NULL
GROUP BY B1.id) relVarTable1
ON relVarTable0.id = relVarTable1.id
LEFT JOIN
(SELECT B1.id AS id, B1.Y AS val
FROM B B1
WHERE B1.Y IS NOT NULL ) relVarTable2
ON relVarTable0.id = relVarTable2.id
LEFT JOIN
(SELECT B1.id AS id, AVG(A1.W) AS val
FROM B B1, AB AB1, A A1
WHERE B1.id = AB1.B_id AND A1.id = AB1.A_id AND A1.W IS NOT NULL
GROUP BY B1.id) relVarTable3
ON relVarTable0.id = relVarTable3.id
LEFT JOIN
(SELECT B1.id AS id, AVG(C1.Z) AS val
FROM B B1, BC BC1, C C1
WHERE B1.id = BC1.B_id AND C1.id = BC1.C_id AND C1.Z IS NOT NULL
GROUP BY B1.id) relVarTable4
ON relVarTable0.id = relVarTable4.id;
'''
conn = pyodbc.connect('DSN=vertica_kdl_dsn;UserName=__;Password=__)
curs = conn.cursor()
curs.execute('SET search_path TO "$user", rcd_test, public, v_catalog, v_monitor, v_internal;')
curs.execute(query)
curs.commit() # comment out to get no error
rows = curs.fetchall() # if above commit called: pyodbc.Error: ('HY000', '[HY000] [Vertica][VerticaDSII] (20) An error occurred during query execution: (20) (SQLFetch)')
curs.close()
fyi我在Vertica社区论坛上找到了一个简单的解决方案:
在dsn [vertica_kdl_dsn]下的odbc.ini中添加resultbuffersize = 0
(HY000不是一个特别有用的错误消息!)