Hunk - SimpleCSVRecordReader

[tt1]

Hi,

I have been recommended the SimpleCSVRecordReader to convert CSV to JSON.

I have tried to set up a Virtual Index Provider, but taking the default settings, but including the following line;

vix.splunk.search.recordreader com.splunk.mr.input.SimpleCSVRecordReader

When I try to use this against a file containing csv data it returns no results, yet by reseting the above line to default it returns results.

Any help appreciated.

[Ledion_Bitincka · Accepted solution]

Was that a typo or did you properly set the record reader as follows?

vix.splunk.search.recordreader = com.splunk.mr.input.SimpleCSVRecordReader

[tt1]

Ledion, many thanks for your help, you reminding me of the search.log helped. After checking that and googling the error in that log, I found an answer suggesting that I had chosen the incorrect version of Hadoop in my provider settings.

Selecting the correct version has solved the problem.

Thanks again for answering.

[Ledion_Bitincka]

The SimpleCSVRecordReader will only recognize certain file extension by default, you can override the default by setting

vix.splunk.search.recordreader.csv.regex

If that doesn't work, it would be great if you provide a link to the contents of search.log for your search so I can help troubleshoot the issue further

[tt1]

Thanks for answering.

Yes using the web interface I have set vix.splunk.search.recordreader to com.splunk.mr.input.SimpleCSVRecordReader

Also tried

com.splunk.mr.input.SimpleCSVRecordReader, com.splunk.mr.input.ValueAvroRecordReader

as per this link

http://blogs.splunk.com/2013/09/27/hunk-setup-using-hortonworks-hadoop-sandbox/

not a lot out there on this subject.