We need some way (unless one exists that I am unaware of - beyond disabling all but the Container I want to run) to fire off containers in particular order. Run Container "Step1" then Run Container "Step2" and so on.
I have heard, or seen, or read (don't recall) some roundabout left-handed way of doing it. Nothing that seemed straight forward to me. I've ended up using 'Conditional Runner' which we just call 'Running Man' (not an official tool - I don't think) of firing off entire workflows in order. So each container would be it's own workflow. You'd need the CReW Macros tab.
Why is this not been addressed Alteryx!! such a simple thing to do... Especially in cases when you need to write to multiple tabs within the same excel file...
I have found that alteryx follows the Input ID order for almost all instances, but there are exceptions (stand along macros - no inputs/outputs - run first)
we have many processed that pull data from 4 sources, output to a file, then do something with that file as input--> server/tableau
what we found is that any time we add a new input, we have to cut and paste the secondary input as the process will run that before the initial output is ran because the new input has a higher ID...
this works as a kludge, but custom order via containers or something would be awesome
First, off I have liked all three ideas to help create any additional traction necessary to make this change a reality.
Second, I helped a colleague recently who asked about this topic.
I'm curious what pitfalls either of you see that might occur here.
The use cases I have tested are limited. My initial thought was to determine a way to game the system and connect multiple containers across the workflow. But then I decided to pass more meaningful data like run time stats, record counts etc.
So take the Container titled "Container 2", the block until done tool prioritizes the output tool first and then generates the single row header stats to be passed to the container titled "Container 4". Container 4 uses a select tool immediately following the append tool to either rename the container 1 stat fields or deselect those columns if not needed. Those stats could be utilized to serve as a DQ check to ensure that the End time of Container 1 is always less than the Start time of Container 2.
My initial concern with this approach was that the Input Tool in the right container above might kick off since it's not directly tied to the Block until done tool. But no matter how long the left container took to process (E.g. 5 seconds or 1 minute) the preceding containers worked as expected. In the example above, the output tool generates an Alteryx database file that is used as the input file in the second container. I updated the left container's input tool with varying number of records to check that the right population was passed. I do realize this would be tedious if you had say 100 containers, but I feel like most times this happens the frequency is far less.
@SeanAdams - Admittedly, I had not read your idea you posted back in 2017. I guess that's the pitfall. I guess you could keep daisy chaining block until done tools to handle additional dependencies but either way it lacks the elegance of a single delivered tool that provides this functionality without all the extra tools.