Splunk Search

## How to chart each data point by another field without any stats function transformation?

Communicator

I think I'm having a brain fart. I want to chart each data point by an address. I don't want (avg, sum, max, min, etc). I want a point on a chart for each dataset. What command am I'm missing? I tried | chart values(DATA) by address ...and all I get is all the data in a block per address.

Thanks,

Tags (3)
1 Solution
SplunkTrust

If you really want a point for each event then you can do something like this:

``````base search | xyseries _time address data
``````

I think that's an ugly approach though, mostly because that does not give you a linear time x-axis. Silence in your data disappears in the chart instead of displaying space for that silence.

Do you really need a point for each event? Maybe you just want more points than you get by default (up to 100), consider increasing that like this:

``````base search | timechart bins=500 avg(data) by address
``````

Interested in high values, but don't want to be misled by single huge maximums? Consider using a percentile function such as `p95(data)` that gives you the 95th percentile value for each time bucket, an often more useful "high" value than a simple `max()`.

If you have drilldowns into single addresses you could consider putting `max(data) avg(data) min(data)` into one chart... just don't do that `by address` for many addresses though, the chart would get very messy.

SplunkTrust

If you really want a point for each event then you can do something like this:

``````base search | xyseries _time address data
``````

I think that's an ugly approach though, mostly because that does not give you a linear time x-axis. Silence in your data disappears in the chart instead of displaying space for that silence.

Do you really need a point for each event? Maybe you just want more points than you get by default (up to 100), consider increasing that like this:

``````base search | timechart bins=500 avg(data) by address
``````

Interested in high values, but don't want to be misled by single huge maximums? Consider using a percentile function such as `p95(data)` that gives you the 95th percentile value for each time bucket, an often more useful "high" value than a simple `max()`.

If you have drilldowns into single addresses you could consider putting `max(data) avg(data) min(data)` into one chart... just don't do that `by address` for many addresses though, the chart would get very messy.

Communicator

Thanks.. The xyseries is just what I was looking for... The timechart or chart means I still have to look at my data as average, max, etc and I wanted the actual value and nothing else.

Communicator

Each event is timestamped. Specifically it's the "oddsets" from an NTP log. I want my chart to plot a point for each offset by the address. So that over time you can see the offset changes for each address that is reported.

SplunkTrust

What do your events look like, and what should the resulting chart look like?

Take the 2021 Splunk Career Survey

### Help us learn about how Splunk has impacted your career by taking the 2021 Splunk Career Survey.Earn \$50 in Amazon cash! Full Details! >

Get Updates on the Splunk Community!