Re: SQL*Loader and csv files

From: (wrong string) öström <lasse.sjostrom_at_telia.com>
Date: Sat, 13 Nov 1999 12:39:11 GMT
Message-ID: <PZcX3.1707$4Y8.200631808_at_newsb.telia.net>


There is a kehy word SEPARATED BY or something like that, you need to specify that key word in the CTL-file. And you might need the key word for TRAILING NULLCOLS I guess it is. Check the doc for the correct words. There are a lot of good examples in the doc file.

Lars

Mike Roca, Jr. skrev i meddelandet <382C7B1A.31BC528C_at_ultranet.com>...
>Hi,
>
>I'm attempting to get SQL*Loader to transfer a csv file into an Oracle
>database, however, I seem to be running into a bit of a problem.
>
>I tried the following command:
>sqlldr LOG=check.log, DATA=nscp.csv, USERID=vlad/math, CONTROL=nscp.ctl
>
>Here's ctl file:
>LOAD DATA
> INFILE 'nscp.csv'
> INTO TABLE nscp
> (dat date "MM/DD/YY",
> cola float external,
> colb float external,
> colc float external,
> cold float external,
> larg integer external)
>
>The CSV file that I am trying to load into Oracle contains over 200 rows
>similar to the following:
>1/12/99,64.500,71.500,64.000,71.063,6625200
>1/11/99,68.938,70.500,64.500,64.500,6818800
>1/8/99,61.938,65.563,61.875,64.750,4272300
>
>I'd concider using the position keyword, but the variables are separated
>by commas rather than tabs, so the positioning is different for every
>row.
>
>-Mike
Received on Sat Nov 13 1999 - 13:39:11 CET

Original text of this message