CSV reporting needs to handle empty reports better.

[elasticmachine]

Original comment by @bhavyarm:

Kibana version: 6.0.0 BC1

Elasticsearch version: 6.0.0 BC1

Server OS version: darwin_x86_64

Browser version: Chrome latest

Browser OS version: OS X

Original Kibana install method (e.g. download page, yum, from source, etc.): from staging

Original X-Pack install method (e.g. by name, offline install, from source, etc.): from staging

Description of the problem including expected versus actual behavior: If user prints csv report on empty results right now- Kibana just prints an empty CSV. User will have to download the csv to realize that it's empty.

Logging this here after discussing it with @kobelb so he can figure out how to handle it.

Steps to reproduce:

1. Open a indexpattern in discover with no results
2. save it and print a report
3. Download the csv - its empty

Example report:
LINK REDACTED

[elasticmachine]

Original comment by @kobelb:

@bhavyarm I removed the bug label since it's technically worked as designed right now, but we could definitely improve on this experience.

[kertal]

To improve the user experience the number of rows a report contains should be displayed. Additionally I think the size of the report would also be helpful. I've added a rows field in the createGenerateCsv function, and with some more adaptions I've got a result, that i've wanted to share and discuss, before i'll continue implementing this approach

If the number is 0 the download button should be disabled. So, what do you think?

[elasticmachine]

Pinging @elastic/kibana-reporting-services (Team:Reporting Services)

[elasticmachine]

Pinging @elastic/kibana-app-services (Team:AppServices)

[tsullivan]

experience the number of rows a report contains should be displayed

Great idea! Working on this in #88303

[tsullivan]

If the number is 0 the download button should be disabled. So, what do you think?

I think the behavior has changed quite a bit since this bug was filed. Today, an empty CSV still contains the search fields as header names, and I think that is useful.

[kertal]

@tsullivan yes, if somebody want's to download suche a CSV, it should be possible (BTW. this was the last issue I kind of contributed to before the elastic recruiter called me, maybe he was missing that feature 😄 )

[tsullivan]

#88303 adds logging about the number of rows are generated from the data, which I think is adequate.

[tsullivan]

Closing this since we add the column headers to the CSV (if they are defined in the job params, and are not just *), and we log how much data is created in the job during execution.