I am using Hunk 6.2.1 and I have some csv format data saved in my hadoop cluster which doesn't have csv header.
By default it seems like Hunk is recognizing the first line of my csv format data as header and not data.
Is there a way to search this kind of data so Hunk doesn't think that the first line content is csv header ?
By default, Hunk will use SimpleCSVRecordReader reader to process csv files and convert csv to json.
By using SplunkLineRecordReader, you can avoid this conversion which enables you to use Splunk Enterprise's search time field extractions settings.
So you will need to use the default line reader to let the splunk parsing pipelines do all the event breaking etc.
To use the default line reader you will need to add "vix.input.1.recordreader = SplunkLineRecordReader" in you virtual index.
[sampledata]
vix.input.1.path = /data/...
vix.provider = cdh4
vix.input.1.recordreader = SplunkLineRecordReader
By default, Hunk will use SimpleCSVRecordReader reader to process csv files and convert csv to json.
By using SplunkLineRecordReader, you can avoid this conversion which enables you to use Splunk Enterprise's search time field extractions settings.
So you will need to use the default line reader to let the splunk parsing pipelines do all the event breaking etc.
To use the default line reader you will need to add "vix.input.1.recordreader = SplunkLineRecordReader" in you virtual index.
[sampledata]
vix.input.1.path = /data/...
vix.provider = cdh4
vix.input.1.recordreader = SplunkLineRecordReader
Thanks SO much for this post. I can't find SplunkLineRecordReader documented anywhere.
We had non-compressed CSV data that was comma separated. When I used SimpleCSVRecordReader it did indeed treat the data as JSON. Not sure why? And this was bad as I was getting pairs of data.
The solution was vix.input.1.recordreader = SplunkLineRecordReader
as listed above.
Thanks again.