DSXchange

Posted: **Fri Sep 22, 2017 5:40 am**

Is it possible to run analysis against a text file? I have an ODBC text file driver setup and imported into IMAM, identified the file, and then imported its metadata into IMAM, then ran quality analysis from the IA thin client. It failed and I found the following fatal message in the corresponding DataStage job log:

Code: Select all

pxbridge(0): ODBC function "SQLPrepare" reported:  SQLSTATE = 42S02: Native Error Code = 1,509: Msg = [IBM(DataDirect OEM)][ODBC Text driver]File not found '/path/project/source/`file.txt'. (CC_OdbcDBStatement::dbsPrepare, file CC_OdbcDBStatement.cpp, line 362)

I noticed a tick mark was introduced between the path and file name. Not sure where that came from, but I was able to View Data Sample OK from the IA workbench. Column analysis run from the workbench fails with the same error. I have opened a PMR.

Posted: **Fri Sep 22, 2017 6:12 pm**

Is this reproducible? Is the character always a back-tick? Any idea where it comes from? Are your ODBC drivers up to date?

Posted: **Mon Sep 25, 2017 7:10 am**

No idea where the back-tick came from. It happens consistently.

As far as I know, we are using the ODBC drivers that shipped with 11.5.0.2 but I will need to double check to be sure. Definitely not using drivers older than that...

I am able to "View Data Sample" on this file from the IA workbench with no issues.

/path/project/source is not the actual path. I have to substitute generic values in place of real values when posting online.

Here is the .odbc.ini header section entry:

Code: Select all

MYPOC=DataDirect TextFile(*.*) Driver

Here is the .odbc.ini driver section entry (it's got mostly default settings):

Code: Select all

[MYPOC]
Driver=/opt/IBM/InformationServer/Server/branded_odbc/lib/VMtxt00.so 
Description=DataDirect TextFile(*.*) Driver 
AllowUpdateAndDelete=0
ApplicationUsingThreads=1 
CacheSize=4 
CenturyBoundary=20
Database=/path/project/source
DataFileExtension=TXT 
DecimalSymbol=. 
Delimiter=
FileOpenCache=0 
FirstLineNames=0
IntlSort=0
ScanRows=25 
TableType=Comma 
UndefinedTable=GUESS

Here is the /path/project/source/QETXT.INI file that was automatically built:

Code: Select all

[Defined Tables]
file.txt=file.txt 
[file.txt]
FILE=file.txt 
FLN=0 
DC=|
TT=CHARACTER
Charset=ANSI
DS=.
FIELD1=colA,VARCHAR,255,0,255,0,

DSXchange

IA quality analysis on a text file?

IA quality analysis on a text file?