Kibana: Data Table and Line Chart have different numbers

[mohd.faraaz]

Hi,

Reading from the same index, if we plot a line chart or create data table across same dimension, we get different results.
The numbers in data table are always less than what is shown in line chart.

Any fixes?

Thanks in advance
Cheers, Mohammad Faraaz

[Joe_Fleming]

There's not enough information here to be able to answer that. How are you building the aggregations for each of the visualizations? Are you applying filters to either of them? Are they the same filters? Some screenshots from Kibana of the two visualizations loaded in Visualize would be helpful, if you can can provide them.

[mohd.faraaz]

Hey Joe,
We are building aggregations for both data table and visualization from the same index and apply same filters to each of them. Then they are placed in the dashboard.
The aggregated count of observations are different in data table and visualization.

Looking for your help.

Cheers, Faraaz

[mohd.faraaz]

Gentle Reminder

[mohd.faraaz]

Hi,
In case more information is required, please let me know.

Regards
Faraaz

[Joe_Fleming]

I tried this locally, and I found that the table vis wasn't showing the correct count, seemingly because it was not preserving the filter. It looks like that's been fixed in newer versions of Kibana though. I can't find the associated pull request so I don't know what version it was fixed in.

What version are you running? And can you try upgrading to the latest version?

[mohd.faraaz]

Hi Joe,
Thanks for simulating this exercise.

We are using Elastic Search 5.0.1 and Kibana 5.0.1 version in our case.

Cheers, Faraaz

[Joe_Fleming]

Hrm, 5.0.1 doesn't seem the have the bug I was seeing.

You say "The numbers in data table are always less than what is shown in line chart," how far off are the numbers you are seeing?

[mohd.faraaz]

Hi Joe,

The tabular data is ~18% lesser than what is represented in visualizations.

Regards
Faraaz

[Joe_Fleming]

I was thinking it's possible that the time buckets were different, which is why the numbers would be off, but that shouldn't really happen, and would probably only be noticeable if you had very high document counts.

A screenshot of the table and the visualization would be really helpful here. It's hard to figure out what might be going on here without looking at it.