This is happening to me too. Even in the output node it shows my 14 fields and 13 rows, but when I run the flow it still gives me "System error: null" with no further explanation.
I am getting the same "system error: null"
I right clicked the step before my output and chose to "preview in tableau" and the data exported to tableau fine. I then exported the data to csv from Tableau to continue my work.
The other two outputs in the flow seem to work fine.
Any idea what's going on?
Same issue for me as well. My flow is similar to Martin's, but with a union as the second to last step. Output steps placed before the union run fine, it's only after the union where the error occurs. I have tried this in versions 2018.1.2, 2018.2.3, and 2018.3.1 with the same result. Hopefully previewing in Desktop and exporting from there will work for now.
I have brought this to the attention of the team at Tableau - could you please attach a twbx workbook with the data included so they can see what is going on-
If you are concerned about confidentiality you can send it to me at Marketanalyticsllc@gmail.com or open a support ticket
I have an issue with a system error null too.
For me, it was a SQL syntax issue, but the crux is a critical Tableau bug:
I have a "Where DataDate >= '2018-01-01'" condition in my Custom SQL and Prep adds an extra ">" in there. Not matter what I do to remove it, it reappears.
There are 3 things with this:
- The critical bug of extra syntax in my custom SQL query
- The fact that the syntax error doesn't trigger a warning at the input step
- The fact that the error message is utterly uninformative.
I sent a ticket already.
Hi Jim Dehner - thanks alot for the support!
Have upgraded to 2018.3.1 and tried it once again, but failed.
I tried to pull the extract from prep while connected to postgresql and it didn't worked, just gave the flow within the packaged file - attached.
Gave another test by:
- first extracting the data from postgres in hyper format; (using prep)
- connected the hyper file back again to prep;
- ran the cleaning steps;
- it worked this time and gave an output file - attached. It also gave packaged flow file - attached.
Haven't raised the ticket yet, but will do now. Any further suggestions?
Don't have a good answer - suggest you raise a ticket on it
In 2018.3.1 I'm experiancing the same problem. Last Friday I made a flow with an output-step 'save to file'. I ran it multiple times without problems.
Today (monday) I started the flow and it gave me the same error: System error: Null
I did some testing and here are my findings
- The flow works fine EXCEPT the output-step
- CHANGE the name of the outputfile (add a 2) and it runs perfectly again
So i tried changing the name back again (remove the 2). But then I get the same error again.
Then it hit me...... The file is in use by Tableau Desktop. That application is open with that file (live connection) in use.
When I close Tableau Desktop en rerun the file in Tableau Prep........ Bingo! Runs perfectly again.
Thanks Matthijs! Shall keep that in mind as well!
2 of 2 people found this helpful
Hi - just wanted to update that finally got a solution! Thanks to the Tableau support perseverance!
The problem was due to INT-SPLIT so was suggested to change to SPLIT only. There were a couple of more steps which made it work; which are:
- Split it by: SPLIT( [Area], " ", 1 )
- Do the clean steps: remove letters, remove punctuation and trim spaces
- Then group and replace the null (or blanks) with “NULL”
- Here is the trick: move the 'group and replace' step to right under SPLIT step
- Duplicate the Area column by calculated field and name it Area Name
- Do the clean steps on Area Name column: remove numbers and trim spaces
- Run the flow!
The PostgreSQL connection doesn't hold null values in string format after cleaning steps. So had to replace it with "Null" or "NA" and move it up; and it worked!
Below is the snip of the steps.
Thanks all for the support!
I had the same problem today. I had to delete my workflow piece by piece until I found the culprit. Then built it out again the same way with no error.
I found if you use field that was transformed to STR (original field is INT) in JOINS output just after JOIN stopped working.
After removing INT->STR transformation - the whole flow is started working.