Unfortunately I have no control over the log data formatting...
it is in format: Field1=Value1|Field2=Value2| ... |Criteria=one,two,three,99.0|...
I have one field, Criteria, that has many values with embedded commas.
Splunk search only give me the first value... I want all values treated as one in a stats count by
I tried below to rewrite them, and do see the changes, but stats still getting only first value.
index=myidx Msg=mymsg | rex mode=sed field=_raw "s/,/-/g"
| bucket span=1d _time as ts
| eval ts=strftime(ts,"%Y-%m-%d")
| stats count by ts Criteria
sorry but I am still only getting the first value before the first comma
do I need to do a re eval
should I be using mvindex in some form?
sorry but I am still only getting the first value before the comma/hyphen.
do I need to re eval the whole line?
Let's go back to your data. The illustrated format, is it raw event or is it from one field? What is the command that you last tried?
I added: | kv pairdelim="|"
but stats is still only showing me the first value before the comma
or - if i keep the rex mode=sed field=_raw "s/,/-/g"
Sorry for misleading you about kvdelim. You still need it when the string is not quoted.
| makeresults | fields - _time | eval _raw = "Field1=Value1|Field2=Value2| ... |Criteria=one,two,three,99.0|..." | kv pairdelim="|" kvdelim="="
Criteria | Field1 | Field2 | _raw |
one,two,three,99.0 | Value1 | Value2 | Field1=Value1|Field2=Value2| ... |Criteria=one,two,three,99.0|... |