在python subprocess.call（）中使用sqlldr失败，但是没有信息

Question

I'm in need of some assistance. I'm attempting to perform SQLLDR from within python. The best method I found was to use subprocess.call. Using the params for another function, I duplicated it within this code. When I run it, I get the appropriate fields, as expected. But, the process returns a 1, which is a failure. I have no additional information and can't locate what could be the problem.

I have verified the data.csv loads into my table from BASH, however python doesn't.

def load_raw():
    DATA_FILE='data.csv'
    CONTROL_FILE='raw_table.ctl'
    LOG_FILE='logfile.log'
    BAD_FILE='badfile.log'
    DISCARD_FILE='discard.log'
    connect_string = os.environ['CONNECT_STRING']

    sqlldr_parms='rows=1000 readsize=50000 direct=true columnarrayrows=100 bindsize=500000 streamsize=500000 silent=(HEADER,FEEDBACK)'

    parms = {}
    parms['userid'] = connect_string
    parms['sqlldr'] = sqlldr_parms
    parms['data'] = DATA_FILE
    parms['control'] = CONTROL_FILE
    parms['log'] = LOG_FILE
    parms['bad'] = BAD_FILE
    parms['discard'] = DISCARD_FILE

    cmd = "userid=%(userid)s %(sqlldr)s data=%(data)s control=%(control)s log=%(log)s bad=%(bad)s discard=%(discard)s" % parms

    print "cmd is: %s" % cmd

    with open('/opt/app/workload/bfapi/bin/stdout.txt', 'wb') as out:
      process = call(cmd, shell=True, stdout=out, stderr=out)

    print process

cmd is: sqlldr userid=usr/pass rows=1000 readsize=50000 direct=true columnarrayrows=100 bindsize=500000 streamsize=500000 silent=(HEADER,FEEDBACK) data=data.csv control=raw_table.ctl log=logfile.log bad=badfile.log discard=discard.log

process returns 1 The log files for log, bad and discard are not created

stdout.txt contains /bin/sh: -c: line 0: syntax error near unexpected token (' /bin/sh: -c: line 0: sqlldr userid=usr/pass rows=1000 readsize=50000 direct=true columnarrayrows=100 bindsize=500000 streamsize=500000 silent=(HEADER,FEEDBACK) data=data.csv control=raw_table.ctl log=logfile.log bad=badfile.log discard=discard.log'

data.csv contains

id~name~createdby~createddate~modifiedby~modifieddate
6~mark~margaret~"19-OCT-16 01.03.23.966000 PM"~kyle~"21-OCT-16 03.11.22.256000 PM"
8~jill~margaret~"27-AUG-16 12.10.12.214000 PM"~kyle~"21-OCT-16 04.16.01.171000 PM"

raw_table.ctl

OPTIONS ( SKIP=1)
LOAD DATA
CHARACTERSET UTF8
INTO TABLE RAW_TABLE
FIELDS TERMINATED BY '~' OPTIONALLY ENCLOSED BY '"' TRAILING NULLCOLS
(ID,
NAME,
CREATED_BY,
CREATED_DATETIME TIMESTAMP,
MODIFIED_BY,
MODIFIED_DATETIME TIMESTAMP)

Answer 1

The error was caused by the silent param. Wrapping it in single quotes allowed the code to work, as here: silent='(HEADER,FEEDBACK)'