Splunk Search

eval - inconsistent math result

splunk_zen
Builder

How is it possible that an eval expression which its components total weight is 100, breach that expected value?

For a given period, I'm getting a result of 101,

while expression components weights are (it totals to 100)

100*(0.08 + 0.03 + 0.12 + 0.16 + 0.06 + 0.18 + 0.06 + 0.06 + 0.25)

and there is not even one component value which exceeds 1.

| eval TimeoutScore=100*( 
0.08*(bw1sdpSdfPccCsApdex) 
+ 0.03*bw1sdpSdfPccMnCsApdex 
+ 0.12*bw1sdpSdfPccPsApdex 
+ 0.16*(0.5*bw1sdpSdfPpsCsApdex +0.5*bw2sdpSdfPpsCsApdex) 
+ 0.06*(0.5*bw1sdpSdfPpsMnCsApdex +0.5*bw2sdpSdfPpsMnCsApdex)
+ 0.18*(0.5*bw1sdpSdfPpsPsApdex + 0.5*bw2sdpSdfPpsPsApdex)
+ 0.06*(0.5*bw1sdpSdfPpsVisaCsApdex + 0.5*bw2sdpSdfPpsVisaCsApdex) 
+ 0.06*bw1careSdfMnApdex 
+ 0.25*(0.5*bw1sdpSdfSlrApdex + 0.5*bw2sdpSdfSlrApdex))
| stats sum(TimeoutScore) as TimeoutScore sum(bw1sdpSdfPccCsApdex) sum(bw1sdpSdfPccMnCsApdex)     sum(bw1sdpSdfPccPsApdex) sum(bw1sdpSdfPpsCsApdex) sum(bw2sdpSdfPpsCsApdex) sum(bw1sdpSdfPpsMnCsApdex) sum(bw2sdpSdfPpsMnCsApdex) sum(bw1sdpSdfPpsPsApdex) sum(bw2sdpSdfPpsPsApdex) sum(bw1sdpSdfPpsVisaCsApdex) sum(bw2sdpSdfPpsVisaCsApdex) sum(bw1careSdfMnApdex) sum(bw1sdpSdfSlrApdex) sum(bw2sdpSdfSlrApdex)


TimeoutScore    sum(bw1sdpSdfPccCsApdex)    sum(bw1sdpSdfPccMnCsApdex)  sum(bw1sdpSdfPccPsApdex)    sum(bw1sdpSdfPpsCsApdex)    sum(bw2sdpSdfPpsCsApdex)    sum(bw1sdpSdfPpsMnCsApdex)  sum(bw2sdpSdfPpsMnCsApdex)  sum(bw1sdpSdfPpsPsApdex)    sum(bw2sdpSdfPpsPsApdex)    sum(bw1sdpSdfPpsVisaCsApdex)    sum(bw2sdpSdfPpsVisaCsApdex)    sum(bw1careSdfMnApdex)  sum(bw1sdpSdfSlrApdex)  sum(bw2sdpSdfSlrApdex)
101 1   0.854839    0.972727    1   1   1   1   0.975410    1   1   1   0.983871    1   1

EDIT:
Thanks to kristian.kolb suggestion, this issue was solved replacing

| eval TimeoutScore=...

by

| eval TimeoutScore=exact(...)
Tags (2)
0 Karma
1 Solution

kristian_kolb
Ultra Champion

splunk_zen
Builder

Dammit... Forgot that could be the culprit.
If you promote the comment to an answer I will gladly accept it 🙂

0 Karma

kristian_kolb
Ultra Champion

Rounding errors? perhaps the exact() function for eval might help?

http://docs.splunk.com/Documentation/Splunk/latest/SearchReference/CommonEvalFunctions

Got questions? Get answers!

Join the Splunk Community Slack to learn, troubleshoot, and make connections with fellow Splunk practitioners in real time!

Meet up IRL or virtually!

Join Splunk User Groups to connect and learn in-person by region or remotely by topic or industry.

Get Updates on the Splunk Community!

Make Your Machine Data AI-Ready in Splunk

Every team is getting ready for AI agents.    Watch On-Demand   In this Tech Talk, Splunk platform specialists ...

How to Break a Kubernetes Pod on Purpose and Let the AI Troubleshooting Agent Find ...

Let's say I've just rolled out a Kubernetes Deployment with a typo in the image tag. Nothing explodes. The pod ...

Splunk Add-on for Microsoft Windows 11.0.3: Parse Once, Search Faster

The updated Splunk Add-on for Microsoft Windows helps customers search Windows events faster by reducing ...