Question Tableau Extraction Issue with Salesforce WebAPI

z_cook82

Prominent
Nov 19, 2018
11
0
510
I am having an issue when trying to extract a large database through Tableau. I have it hooked up to the Salesforce WebAPI and am using it through my System Administrator account. Whenever I try to extract my large Contacts database, it continuously fails with a generic error "Tableau failed to create an extract. This data source requires an extract before you can start your analysis."

I have gone through Tableau support and they say it has nothing to do with Tableau and it's on the Salesforce side, saying that the tables are unable to be extracted. I talked with Salesforce support, went through everything with them, and still have the same issue. Both supports are pretty much useless at this point and my client and I are both stumped. Anyone have any ideas?

Currently, we've verified every field in the Contacts table has Read Access and everything that can has Edit Access. The Contacts database has 297 fields in it. I'm not sure where to go from here, any help would be appreciated.

Thanks!
 

Ralston18

Titan
Moderator
Which Tableau product: Tableau Data Management?

An "extract"?

Found this:

"If the data source is an Extract, all refreshes of the extract are managed by Tableau Server and can only be refreshed by the server. For more information, see Best Practices for Published Data Sources"

And that led to the following link:

https://community.tableau.com/thread/280099


Did anyone or was anyone at Tableau able to truly explain what the meaning of "Tableau failed to create an extract. This data source requires an extract before you can start your analysis." What that error really means and what can cause that failure?

It is their product and they should be able to tell you at least that much....

Sans blaming Salesforce and/or you.

I found the following link:

https://onlinehelp.tableau.com/current/pro/desktop/en-us/refreshing_data.htm

Note the orange "Extract Your Data" link under the Extract Section.

My suggestion is to get every one involved in a video conference (or a least a teleconference) and work through the extraction process step by step with everyone watching and involved.

No finger pointing or blaming allowed. Just get each party to "sign-off" on each step or point out some issue, misconfiguration, skipped step, whatever....

With any luck there will be an "Eureka" moment.
 

z_cook82

Prominent
Nov 19, 2018
11
0
510
Which Tableau product: Tableau Data Management?

An "extract"?

Found this:

"If the data source is an Extract, all refreshes of the extract are managed by Tableau Server and can only be refreshed by the server. For more information, see Best Practices for Published Data Sources"

And that led to the following link:

https://community.tableau.com/thread/280099


Did anyone or was anyone at Tableau able to truly explain what the meaning of "Tableau failed to create an extract. This data source requires an extract before you can start your analysis." What that error really means and what can cause that failure?

It is their product and they should be able to tell you at least that much....

Sans blaming Salesforce and/or you.

I found the following link:

https://onlinehelp.tableau.com/current/pro/desktop/en-us/refreshing_data.htm

Note the orange "Extract Your Data" link under the Extract Section.

My suggestion is to get every one involved in a video conference (or a least a teleconference) and work through the extraction process step by step with everyone watching and involved.

No finger pointing or blaming allowed. Just get each party to "sign-off" on each step or point out some issue, misconfiguration, skipped step, whatever....

With any luck there will be an "Eureka" moment.
So, I totally forgot to put it in there, but the client only has Tableau Desktop, they do not have Tableau Server. They were able to get Tableau Desktop due to them being a Non-Profit and were able to take advantage of one of Tableau's offers for being a Non-Profit.

Tableau support pretty much just blew me off and put all blame on Salesforce and that was that. I spent a good 2 hours or so with the Tableau support (all with the client using the application, and using the companies billed time) only to tell me "sorry, looks like Salesforce can't extract their tables. Not our issue, go talk to Salesforce support"and then 10 minutes later I got an email saying the case was being marked 'Closed'
 

Ralston18

Titan
Moderator
My suggestion is that you and your client call Tableau back and escalate the issue.

No point in offering the product to a Non-Profit and not providing at least some support.

Tableau should be able to pinpoint the reason why the extraction is failing. What extract criteria is not being met , what could/might be wrong with the source (Salesforce ) data extract? Some mismatched field type....?

Is it viable to attempt some smaller test extractions from the Contacts Database. Idea being to determine if the failure(s) can be narrowed down to certain data fields, sorts, filters, quantifiable factors, etc...?

Start simple and proceed as you deem applicable and appropriate - you know the software, the process, and the data.

(I'm just rooting from the sidelines here....)

The more information you can glean out about the problem then the more everyone can re-focus and go forward.
 

z_cook82

Prominent
Nov 19, 2018
11
0
510
My suggestion is that you and your client call Tableau back and escalate the issue.

No point in offering the product to a Non-Profit and not providing at least some support.

Tableau should be able to pinpoint the reason why the extraction is failing. What extract criteria is not being met , what could/might be wrong with the source (Salesforce ) data extract? Some mismatched field type....?

Is it viable to attempt some smaller test extractions from the Contacts Database. Idea being to determine if the failure(s) can be narrowed down to certain data fields, sorts, filters, quantifiable factors, etc...?

Start simple and proceed as you deem applicable and appropriate - you know the software, the process, and the data.

(I'm just rooting from the sidelines here....)

The more information you can glean out about the problem then the more everyone can re-focus and go forward.
So we've been able to do a test extraction with a minimal amount of data (the table itself in SF has 297 fields, we narrowed it down to about 150 or so) and it was able to extract properly. At this point, Salesforce support seems to be more helpful than Tableau support with Tableau itself haha. Since there are 297 fields in the Contacts database, it will be very time consuming and difficult to kind've pinpoint if any specific fields are failing due to a macro, weird filter, etc.
 

Ralston18

Titan
Moderator
Narrowed down to 150 fields.

Okay -- that is progress...

Narrow the number of field down even further.

Try groups of 50 fields, the 1st 50, the 2nd 50, 3rd 50. (You pick the number of fields)

Or try an extraction with only text fields, then numeric fields, then only date fields, etc..

Key is to be thoughtful and methodical about it all.

Think about the Contacts database: What field(s) are most likely to have "bad data". Maybe some group of characters in one field that comes across as a control code or false delimiter? Macros?

Does take time but sometimes all you can do is grind through it.
 
  • Like
Reactions: Alabalcho

z_cook82

Prominent
Nov 19, 2018
11
0
510
Narrowed down to 150 fields.

Okay -- that is progress...

Narrow the number of field down even further.

Try groups of 50 fields, the 1st 50, the 2nd 50, 3rd 50. (You pick the number of fields)

Or try an extraction with only text fields, then numeric fields, then only date fields, etc..

Key is to be thoughtful and methodical about it all.

Think about the Contacts database: What field(s) are most likely to have "bad data". Maybe some group of characters in one field that comes across as a control code or false delimiter? Macros?

Does take time but sometimes all you can do is grind through it.
I agree, this is the best way to do this, however with the amount of time that we've put into this already (initial troubleshooting, tech support calls, etc) we've got about 6 or 7 hours of billable time into this ticket and we're trying to minimize it at this point before the client decides this is a waste of company billable time. We're able to extract 150 different fields, but the last 147 I don;t think we'll have enough time to justify figuring out which ones work and which don't down to a singular level without adding an additional 2 hours or so, then more time on top of that to diagnose the specific issue with those fields
 

Ralston18

Titan
Moderator
Tableau's Forums might help but doing so will still require someone's time and effort.

https://community.tableau.com/welcome

Or perhaps there is some utility or diagnostic tool that can help: third party or otherwise. (However, I would hope that any such options would have been suggested or offered by now. Via either Tableau or SalesForce,)

Unfortunately time and budget/cost constraints are out of my purview here.

Likewise that includes making any necessary "trade-off" decisions that compare the costs\benefits of not solving the problem versus the costs\benefits of solving the problem.