Best way to index .csv files with some common fields

[simonbuskens]

Hi
I have a series of .csv files (1 for each month) where the first 100 fields are the same, but after that there are about 4 or 5 fields that are specific to that month only.

What is the best way to add that data into Splunk?

Will I be able to add them all to the same index, or do they need their own index (so then I have to search across them all)

[esix_splunk · Accepted solution]

For your sourcetype, define the fieldnames in the header in all of your csv's in props.conf.

Then, define additional fields in that header for the sourcetype. E.g.,

fields = alwaysherefield1, alwaysherefield2, alwaysherefield...30, sometimesherefield31, sometimesherefield32, sometimesherefield33

If those fields dont exists, there wont be a value set. In the CSVs that have the fields, the values will be set. Then in your search, you can rename the fields to whatever you want them to be.

Refer to here for props.conf examples :
http://docs.splunk.com/Documentation/Splunk/6.2.1/Data/Extractfieldsfromfileheadersatindextime

[asifhj]

Files name are same or different? if the file name are different you can put in same index and can write queries according to source field.

[kml_uvce]

you can add in same index and source name will be different and you can search based on source name

[simonbuskens]

won't the source name be the same (if they are saved in the same folder)

[esix_splunk]

Source name will be path+filename, sourcetype will be whatever you defined it to be.

[asifhj]

Nope, folders cannot have two files with same name.