ZEPPELIN-672] Add Datatables instead of normal tables

[ankurmitujjain]

What is this PR for?

It is an extension of #6 #714
It allows user to export data in a paragraph to a TSV/CSV file.

What type of PR is it?

Feature

Todos

• - Improves the current Table features like Search, Fixed Headers, Sorting

Is there a relevant Jira issue?

ZEPPELIN-672

How should this be tested?

1. Create a paragraph with data in %table view
2. Click on TSV/CSV button to export CSV/TSV file

Screenshots (if appropriate)

Questions:

• Does the licenses files need update?
  Need to have MIT license for Datatables.
• Is there breaking changes for older versions?
  No
• Does this needs documentation?
  No

[ankurmitujjain]

@adejanovski @corneadoug @bzz @Leemoonsoo
I am done with integrating datatables, have to update a failing testcase...

@corneadoug can you please let me know any example for Chart Pivoted data to verify if this is completed or not...

Thanks
Ankur

[ankurmitujjain]

@adejanovski @corneadoug @bzz @Leemoonsoo @felixcheung

Hello Team,

Can you please review this PR?

Thanks
Ankur

[corneadoug]

Thanks, I'll test that out

[prabhjyotsingh]

I see following issues:

• There is a horizontal scrollbar
• When i scroll, table data goes above header

[ankurmitujjain]

Thanks for reviewing @prabhjyotsingh
-I will look into horizontal scroll bar

Thanks
Ankur

[ankurmitujjain]

Hi @prabhjyotsingh ,

I had updated code as per your comments please review..

Thanks
Ankur

[prabhjyotsingh]

Tested it with different size of data. Looks good to me.

[ankurmitujjain]

Hi @corneadoug can we merge it ?

[felixcheung]

@hyonaldo @zhongneu you have both created PRs on this before, are you ok with this approach?

[corneadoug]

I found some css problems depending on the paragraph size:

Also, when we scroll, the download buttons and search input can't be seen:

Btw, I saw that the Allowing Export to be not only Table based, but also of Chart Pivoted data task is checked, however I didn't see any button when switching to a graph

[ankurmitujjain]

Hi @corneadoug : Sorry I wrongly marked pivot chart data as complete...
Actually I have to remove that feature for now, I will be exploring on how we can export data from graphs as per you suggestion, but for now I would like to keep this feature away from this PR...

However we can always export graph data in table view.

This PR I had planned only for implementing datatable and its export feature....
Please let me know if you are ok with this...

Regarding CSS problem: I will look into it...

Thanks
Ankur

[jmrr]

@ankurjainyash any progress with this? I'm experiencing some merge conflicts and master is evolving fast.

@Leemoonsoo I've seen that the export capability is clearly set in the roadmap you sent few weeks ago. What do you think it will be the final approach? Any of #725 , #714 , #6 #89 or #672 or a combination of some/all? I've tested #672 and its data display with search/sort seems more complete.

[corneadoug]

I think we would be better off using multiple PRs:

• ZEPPELIN-672] Add Datatables instead of normal tables #761 to replace the table only (if it brings better performances)
• Exporting to TSV #714 to have a download button available at all time for front data
• Added CSV and Tableau export functionality #725 to have a download button available at all time for back-end data

The good thing would be that both back-end and front-end data download would use the same UI, and it wouldn't add additional UI elements in the view

[r-kamath]

#761 to replace the table only (if it brings better performances)

What is the max number of rows this lib can support ?

[ankurmitujjain]

@corneadoug , please also review comments from other users... #6.

[fireboy1919]

See: #822

[corneadoug]

@ankurmitujjain I made a PR to your branch ankurmitujjain#1

[ankurmitujjain]

@fireboy1919 is it possible to download output in the form of CSV/TSV?

[fireboy1919]

Yes, #822 does that. It's purely backend, using resource pools. You might consider it an alternative to using #714.

[ankurmitujjain]

ok if thats the case than I believe datatable dont need download functionality which will be handled by #822 . please correct me if I am wrong...

[ankurmitujjain]

@fireboy1919 is it possible to download both server side data and front end table data?

[fireboy1919]

It's a REST call, so potentially, you could make a client side call to combine the two pieces of information.

But the problem is the same one the backend is solving by being backend - the memory limits imposed by the browser are small compared to what the server can handle.

A smarter approach would probably be to get that kind of thing would be to make a different kind backend component that can transform the data the way that you want it to be transformed.

[ankurmitujjain]

So basically you are suggesting to use rest endpoints in download buttons to download server side data....

[darionyaphet]

hi @ankurjainyash could support a option about save the CSV/TSV to server ? It's very useful .

[ankurmitujjain]

@darionyaphet . If you want to save data as csv on server you can use below code...

spark-csv

 val selectedData = df.select("year", "model")
 selectedData.write
.format("com.databricks.spark.csv")
.option("header", "true")
.save("newcars.csv")

[darionyaphet]

@ankurmitujjain 👍

[ankurmitujjain]

Updated code to fix testcase, please review

[corneadoug]

Merging this

[felixcheung]

Could you check out #858? Seems to overlap with this PR
#858

[corneadoug]

@felixcheung already saw it and I will check it.
However, while we test + discuss the other PR, there is no reason not to merge this one.
If the other is better, then it can simply replace this one

[darionyaphet]

This PR have merge into master branch ? @ankurmitujjain @prabhjyotsingh

[AhyoungRyu]

@darionyaphet Yes it does. See here : d481b56