Skip to main content
  • EN

Hi list.

I have a master workspace in Workbench, which starts with a Creator calling two SQLExecutors, and eventually calls two WorkspaceRunners.

Each SQLExecutor fetches a list of tables in a dataset, adding their feature count (numbers of rows).

I specifically tell FME to process the smaller dataset (i.e. its SQLExecutor) first, but each “thread” also have a (blocking) Sorter transformer (sorting by feature count) before calling (its) WorkspaceRunner.

It seems that the result from the two blocking Sorter transformers are processed by FME in the wrong order regardless of which order the features have arrived.

How can I ensure that the Sorter for the smaller dataset (i.e. VAT above) executes first, before the Sorter for the larger dataset ?

As is seen, I tried to rename the Sorters, so the smaller was alphabetically listed before the larger, but it didn’t change anything. The Logger logged the larger dataset first.

Hi ​@lifalin2016 ,

I think you can remove the Sorters, if you use ORDER BY clause within the SQL statements to sort the output features.

Hi ​@lifalin2016 ,

I think you can remove the Sorters, if you use ORDER BY clause within the SQL statements to sort the output features.

True in this case, but it’s a work-around, not a solution. After posting the question I did in fact do this.

But it might not be possible in all cases, so I’d still like an answer to my question.

Not sure if I understand correct, but right clicking on one of the connections from the Creator will let you set connection runtime order.

Do you perhaps want to determine the order of triggering the two SQLExecutor with some condition or criteria at run-time? If so, can you please explain what the condition is.

Not sure if I understand correct, but right clicking on one of the connections from the Creator will let you set connection runtime order.

Yes, and I did that. But FME ignores that order when passing features on from the blocking Sorter transformers. Why this happens is the essence of my question.

Do you perhaps want to determine the order of triggering the two SQLExecutor with some condition or criteria at run-time? If so, can you please explain what the condition is.

The order is determined by the order of the two connections from Creator.

The lower connection is 1 of 2, and the upper is 2 of 2.

Not sure if I understand correct, but right clicking on one of the connections from the Creator will let you set connection runtime order.

Yes, and I did that. But FME ignores that order when passing features on from the blocking Sorter transformers. Why this happens is the essence of my question.

OK. I was confused by:  

As is seen, I tried to rename the Sorters, so the smaller was alphabetically listed before the larger, but it didn’t change anything. The Logger logged the larger dataset first.

 

But ontopic, I think I have seen this behaviour before. (transformer after sorter ignoring sorted order) In my case it was SQLExecutor, Sorter, PythonCaller where the PythonCaller processed features randomly, not from small to big as I preferred. I worked around it but if I must guess it probably has to do something with the ‘newer’ bulk data handling. This can be tested with a transformer which breaks bulk mode, I think that de Deaggregator does this, but I am not sure.

Community Stats

31,852
Posts
120,988
Replies
39,485
Members

Latest FME

Terms & ConditionsCookie settings
A product of
Safe Software respectfully acknowledges that we live, learn and work on the traditional and unceded territories of the Kwantlen, Katzie, and Semiahmoo First Nations.