New TableRow-To-Variable-Loop-Start node and missing values


I realized that there is a change in the handling of missing values. I would like to understand why missing values are not allowed as flow variables anymore.

In the past, I created several workflows where the loop input table contains some metadata (including missing values). Within the loop, I'm joining back the metadata with a Variable-To-TableColum node. I need to rewrite these workflow parts, as the missing values in the metadata are not transported properly anymore.

Can somebody please explain why this has changed?


Hi Antje, 

In 2.9, we actually added support for missing values in these nodes.  If you configure the nodes, you will see options for how to deal with missing values.  A quick test on my part and this appears to be backward compatible, with th edefault optioin being abort, which is the same behaviour I see in 2.7 and 2.8.  

Is something else going on?





Backward compatibility is fine. But I cannot keep on working like I used to do.

Now, I am forced to choose one of the options 'fail', 'use defaults' or 'omit'. But I don't want to use defaults, I would like to keep it as missing value (and of course, I don't want to choose 'fail' or 'omit'). That was possible before, but not anymore, no?

No this was not possible. At least with my version the node failed with " Execute failed: Missing Values not allowed for Variable Values!"

As I remember it, it always used to fail for me with missing values in previous versions. A real pain it was.

I am very pleased with the new handling of missing value variables.




Hmm, you might be right. Could be that my metadata did not contain missing values...

I'm really sorry - I thought it was possible.

Okay, then I need to think about how to change my workflow to keep the missing values (I might need to join it after the loop but as I have big tables (few millions of lines) this node will take some time...

Any faster suggestion?

The task looks like this: I have a table containing file names and metdata for each file. I need to read in all these files and I want to join the metadata.