adding a new task into the middle of process flow

[cnilsen]

Hi All.

So.. I have this long process flow I'm working on. Prior to the end of it (and its reports) I have to add a new task to do a join with 2 new tables into the process.. so that some of their data can be used for some new calcs in the final query of the process. How can I 'hijack' the process and do so?  I have tried, but then the process branches off and I have to find a way to get the final calcs and queries back in line.

Thanks,

Chris N.

[TomKari]

Really easy, I'm pleased to announce!

1. Set up whatever new nodes you need, and connect them as needed. Don't connect them to anything else.

2. Connect the OUTPUT of your new set of nodes to the required existing sink node. Note, you'll now have the old and the new linking in parallel to your sink node.

3. Depending on whether you need to create an EG link, or it's an implied data link, at this point make whatever kind of connection you need to between your old source node and your new set of nodes. Now you'll have your old source node output split and going to both your new set of nodes and your old sink node.

4. Get rid of the link from your old source node to your old sink node.

Tom

[ChrisHemedinger]

Don't forget that you can easily modify the data reference leading into a task with right-click, Select Input Data.

You can modify data in the Query Builder by selecting Table Properties in the Query Builder interface and browsing to an alternate data table, thus modifying how data flows through the project.

[cnilsen]

Hi Chris,


So by changing the input name of a data string.. I can realign my job stream back into a a single path?


for example.

my process goes from task 'A' to 'B' to 'C'

I do my new table join after 'B' as needed resulting in output 'B2'... and then change the input of 'C' to 'B2'?


Chris N.

[ChrisHemedinger · Accepted solution]

---

@cnilsen wrote:




So by changing the input name of a data string.. I can realign my job stream back into a a single path?

---

If by "changing the input name" you mean "changing the input data reference" from 'B' to your newly created 'B2', then Yes.  You're not just changing a text string (as you might do in a pure code solution), but effectively rearranging your process flow to include this additional intermediate step.  As long as you don't change the attributes of the input data (same column names and types), the downstream processes won't suffer any ill effects.

Chris

[cnilsen]

Got it. Thats what I was hoping for.
Thanks again!

Chris N.