Dashboards & Visualizations

How to create a search results that only takes the monthly data for every other Wednesday?

ichesla1111
Path Finder

Hello!

I am making a search that searches data from Sunday-Tuesday one week and Sunday-Wednesday the next week. 

I have a field called "date_wday" which contains data values representing the day of the week (sunday, monday, tuesday, .....,saturday).

My search: index=blah date_wday=monday OR tuesday OR sunday

My search successfully filters the data to pull from Sunday-Tuesday but how can I also have it add data from every other Wednesday?

ex. Week 1: Sunday-Tuesday
       Week 2: Sunday-Wednesday
       Week 3: Sunday-Tuesday
       ......


Thank you

Tags (2)
0 Karma
1 Solution

bowesmana
SplunkTrust
SplunkTrust

You'll probably have to pull in data for Sun-Wed and then discard the data you don't want. You will have to do a calculation about which week it is, so you won't be able to add simple search criteria.

However, if you search back 2 weeks, it is presumably week 1 Sun-Tue and week 2 is Sun-Wed, but it you search back 3 weeks, does it mean the oldest week is Sun-Tue, 2nd oldest Sun-Wed and most recent Sun-Tue?

If so, you can calculate the age in weeks of each event and then add a where clause to filter out the Wednesday on even weeks, e.g. this example works on the _internal index and will treat the current week as including a wednesday, the previous not. If you search an even number of weeks, it will always give tuesdays on the oldest week and then alternate

 

index=_internal date_wday="sunday" OR date_wday="monday" OR date_wday="tuesday" OR date_wday="wednesday" earliest=-2w@w latest=now
| eval age_in_weeks=floor((now() - _time) / (86400 * 7))
| where age_in_weeks % 2 == 0 OR date_wday!="wednesday"
| bin _time span=1w
| fields _time age_in_weeks date_wday
| stats min(age_in_weeks) as min_age max(age_in_weeks) as max_age values(date_wday) as date_wday count by _time

if you want to make if based on the search window and always exclude wednesday for the oldest week in the search range, you'll need to know the search range, so use 

| addinfo
| eval search_weeks=floor(now() - info_min_time / (86400 * 7)
...

to determine the number of search weeks and then tweak the where clause above as needed

 

 

View solution in original post

ichesla1111
Path Finder

This is awesome, thank you!!! It worked.

0 Karma

ITWhisperer
SplunkTrust
SplunkTrust
index=blah [| makeresults 
    | eval week=mvrange(0,4)
    | mvexpand week
    | eval earliest=relative_time(relative_time(now(),"-1mon@mon+6d"),"+".week."w@w")
    | eval latest=relative_time(earliest,"+".(3+(week%2))."d")
    | fields earliest latest]

bowesmana
SplunkTrust
SplunkTrust

You'll probably have to pull in data for Sun-Wed and then discard the data you don't want. You will have to do a calculation about which week it is, so you won't be able to add simple search criteria.

However, if you search back 2 weeks, it is presumably week 1 Sun-Tue and week 2 is Sun-Wed, but it you search back 3 weeks, does it mean the oldest week is Sun-Tue, 2nd oldest Sun-Wed and most recent Sun-Tue?

If so, you can calculate the age in weeks of each event and then add a where clause to filter out the Wednesday on even weeks, e.g. this example works on the _internal index and will treat the current week as including a wednesday, the previous not. If you search an even number of weeks, it will always give tuesdays on the oldest week and then alternate

 

index=_internal date_wday="sunday" OR date_wday="monday" OR date_wday="tuesday" OR date_wday="wednesday" earliest=-2w@w latest=now
| eval age_in_weeks=floor((now() - _time) / (86400 * 7))
| where age_in_weeks % 2 == 0 OR date_wday!="wednesday"
| bin _time span=1w
| fields _time age_in_weeks date_wday
| stats min(age_in_weeks) as min_age max(age_in_weeks) as max_age values(date_wday) as date_wday count by _time

if you want to make if based on the search window and always exclude wednesday for the oldest week in the search range, you'll need to know the search range, so use 

| addinfo
| eval search_weeks=floor(now() - info_min_time / (86400 * 7)
...

to determine the number of search weeks and then tweak the where clause above as needed

 

 

Get Updates on the Splunk Community!

Splunk Observability for AI

Don’t miss out on an exciting Tech Talk on Splunk Observability for AI! Discover how Splunk’s agentic AI ...

[Puzzles] Solve, Learn, Repeat: Dereferencing XML to Fixed-length events

This challenge was first posted on Slack #puzzles channelFor a previous puzzle, I needed a set of fixed-length ...

Stay Connected: Your Guide to December Tech Talks, Office Hours, and Webinars!

What are Community Office Hours? Community Office Hours is an interactive 60-minute Zoom series where ...