Alteryx Designer Desktop Ideas

The JOIN tool could use some love.  Let's consider merging the JOIN and UNION functions into a single tool.  Instead of strictly L, J, and R outputs, we could have an option to allow for all standard SQL joins:

 

• Cross Join (Warning!!!)
• Inner Join (boring)
• Left Outer Join (saves time configuring Union)
• Right Outer Join (saves time ...)
• Full Outer Join (saves time ...)

Being able to JOIN on case-insensitive values is a big bonus (resisted urge to BOLD and change font size).

Being able to JOIN on date-range is often requested.

Being able to JOIN on numeric-range is often requested.

 

If we are combining tools, getting UNIQUE on L or R (or both) inputs would also save time.  Most JOIN errors are because the incoming (R) data contains duplicates by KEY.

 

cheers,

 

Mark

 

Hi @NicoleJ 

I would like to be able to use the join tool to join on inequalities.  We could join two tables, A and B on A.value is >= B.value1 AND A.value <= B.value2.  This would replicate the "between" function in SQL.  The equvalent feature in Tableau is pictured below.

I've come to realize that the JOIN tool is case-sensitive by design but it would be helpful if you could turn that behavior on/off (via checkbox?) within the JOIN tool.  For those of us that work predominantly in database environments that are not case-sensitive, this default behavior has caused me problems many times.  Having to force the case to either upper or lower upstream of the JOIN on both flows in order to ensure a successful join is an extra step that would not be necessary if you could disable case-sensitive with a checkbox.

We build some pretty robust maps with multiple connections and it would be great to copy the map tool and paste it with all of the connections when we want to tweak the map slightly but keep our original map.  It is a regular occurrence for us to have a very detailed map grouping by trade area name and then may want to have an overview map with all of the same connections but slightly different layout.  Tracking down the connections, reconnecting them and naming them accordingly takes a substantial amount of time even in the most organized of workflows.  This function would be a huge time-saver.  It would also be of value with joins and unions - anywhere you have multiple streams coming in.

In other data programs like access or Toad you can put conditions on a join. You can choose if something is greater than, less than, left join, right join, etc.
But with Alteryx you are only allowed to join a perfect match. It would be really great if you could add that functionality into 9.0

I think it would be nice to be able to more easily reorder fields that you're joining by in the Join tool.

 

 

For example, I have already joined by CASS_Address and CASS_City. After I did this, I realized I wanted to go back and join on Name, too, and I want that to be first. How the tool is configured now, if I want Name to be first, I must redo all of the drop downs. I would like to be able to add Name to the next set of open drop downs then use some arrow buttons to be able to move them up in the order (similar to the Summarize tool).

I think it would be incredibly helpful for Alteryx to include a "Fuzzy Join" operator, similar to what is described in this article: http://www.decisivedata.net/blog/alteryx-fuzzy-join-workflow/

 

Virtually every client/project I work on, there is a nead to clean up data.  Most of the time, that involved standardizing to some existing list of data.  However, as we all know, data from differnet systems or being manually collected will not match perfectly in all cases.  This is most often when I tend to use the Fuzzy Match tool.

 

However, I have to use a lot of weird steps to effectively create a "Fuzzy Join", which is something I've done using database functions in the past.  I think it would be great if a new tool were created that would do the following:

• Accept two inputs, one for the "raw" data and another for the "list" of data to match to.
• Perform a fuzzy join based on similar functionality to the fuzzy match, convert data to metaphone keys and then run Jaro/Levenstein matches.  By default, return only the highest matching result.
• Expand the pre-process functionality to include words to exclude from the analysis (beyond just "and", "the" and "in").  
• Match on the whole string.  No need to try and do joins based on partial words within a string.

 

This seems like a very common thing (I've created a macro for this anyway) that could be made to be simpler for everyday use.

 

Thanks!

 

I'd say that 95.437% of the Joins I do are straight Inner Joins.

 

So each of those times I have to remember to go down to the Select part of the Join tool and deselect all the fields I joined on the Right Side since they'll be duplicates.

 

I'd like a checkbox like below (defaulted to CHECKED)  to deselect all the joined fields from the right hand side. In the rare cases where there's a need I could uncheck it.

 

Deselect R join fields

When bringing data together it is often needed to assign a source to the data.  Generally this happens when you union data and need to know things later about the data for context.  It would save time to generate a source field that is assigned based upon the input connections of the union tool.  Perhaps when unioning data you can assign a name to each input stream?

 

 

I want a feature to enable join by custom conditions. Currently, in Join tool, allowed condition is only equality of specific fields and specific position, however, in SQL, we can join data by much more flexible conditions like;

SELECT TableA.id FROM TableA INNER JOIN TableB ON TableA.id=TableB.id and TableA.value > TableB.value  

Of course, my idea can be easily realized by using combination of Appendix Field + Filter tool, but I meant to say is that Appendix-Fields is quite expensive operation in calculation cost, and it would generate many unnecessary records, which is annoying us in case of handling a huge dataset.

 

I suppose this kind of flexible conditions can be specified by using expression editor, thereby configuration window of this feature would look like the below image; Adding one more radio button option, and expression editor similar to one used in Filter tool.

 

Any positive/negative feedback on my idea would be appreciated. Thank you for your attention!

Alteryx is unlike many BI tools in the sense that it joins NULL. It is difficult to think of another platform that has this behaviour. Either people know about this and work around it or they don't and their joins are a ticking time bomb. Please add a check box to the Join and Join Multiple tools to allow or prevent joining NULL. This will serve to remove the need for workarounds as well as educate users about this default behaviour.

When the append tool detects no records in the source, it throws a warning. I would like to have the ability to supress this warning. In general, all tools should have similar warning/error controls.

Have you ever used a Join tool with several (or many) Join fields, looked at the the L and R outputs and wondered, why didn't these records join? When there are many columns in your data, this can be a hard question to answer. It would be very handy if Alteryx could somehow report the Field(s) that each record failed to join on (perhaps as an optional added field to the L and R outputs).

Back in the days of Visual Basic 6, and Delphi 3 - the development environment for software coding started introducing the idea of Optimization hints - specifically looking at things like variables you'd declared but never used.

 

All of the Alteryx tips guides say "as early as possible use a Select tool to remove fields that are not used in the workflow" - it would be very useful if the Alteryx system did a reverse walk through the workflow (from outputs back to inputs) and suggested fields that were not used anywhere or components that could / should be removed because they don't go anywhere.

 

For example

• Unused Components: if I have a filter with a summarize on both legs, and only one leg goes to an output - then the summarize on the other leg is just a waste of processor and memory.   
• Unused Fields: If I've been lazy and brought in all the fields in a spreadsheet and only used two of them through a complex and heavy workflow - then suggest that I only bring in 2; or that I add a select component for those two fields early on
• Ineffective data types:
  • Numbers stored in a double that could be stored in an int16; 
  • numbers stored in strings;
  • strings stored in very large containers which only need 10 characters;
  • dates stored in strings
  • Using a field for "True" and "False" instead of the Boolean True and False (1 and 0)

By adding a reverse-walk to a workflow (sort of like an A-B-Tree prune on a game-tree) - Alteryx could spot many of these kinds of issues - and by observing the workflow in action over a few iterations, it could identify the remainder.

 

 

Today, any Alteryx tool with "Select" functionality has an option for "Dynamic or Unknown Fields" which, when checked, allows any new fields to pass through that tool.  This is a great function for most of the tools as you can allow workflow updates to pass through the tool without issue.

However, in the Join tool, there are some use cases where there is NEVER a reason to pass new fields from one side or the other into the tool, but you might still want new fields from a primary process.  Examples being something like a lookup/cross-reference to do an inclusive join, where adding new fields to the lookup might inadvertently pass these downstream.  Having the option to only allow unknown fields from one side through would greatly enhance this output.

Our company is still using 9.5 so if this is addressed in 10....I appologize.

 

Currently the Join Tool Options drop down has [Select-->Select All] and [Select-->Deselect All]. I think an additional [Select-->Select All Left] and [Select-->Select All Right] would be handy.

 

Thank You

Here is the issue I have, when you are using a Join tool and you have multiple columns that you are joining on (to the point that they don't all show in the 
Configuration window), i have a tendency to use the mouse scroll wheel to move down to see additional columns i am joining on.  The mouse scroll controls different things depending on where your cursor is.  If your cursor is over the Left or Right columns then the scroll button will change the Fields you are using to join on.  I have messed up more workflows then i care to mention due to this.  I do not think it is appropriate for the scroll wheel to effect and change the fields in the configuration window and it should only be used to scroll up and down in the configuration window.  

 

 

The Find Replace tool has a checkbox to do a case insensitive find. It would be fabulous if the Join and Join Multiple tools had a similar checkbox.

 

I frequently have to create a new field in each data stream, convert the data I want to join on to upper case, perform the join and remove the extra "helper" fields. Using the helper field is needed in my case in order to preserve unique capitalization (i.e., acronyms within the string, etc.). 

Dear Users, Fans, Compatriots, and Fellow Alteryx Nerds:

One of my favourite parts of using Alteryx is that in all the in-memory tools, there is a quick-and-dirty count in each of your tools' output nodes. You know, you use these all the time and when you switch back into SQL, you get frustrated with having to run the query two or three times just to see the count in each of your join outputs. 

One thing I'm missing as an INDB user is that I have to employ a manual workaround to see what is happening. INDB tools are a bit black-box in that we don't see the counts.

All I want...

 

 

I've been using this workaround for a little over a year now and I haven't found it to be incredibly taxing on my resources, so I'm wondering if Alteryx may be able to look into doing this on the back end to make the INDB experience that much closer to the in-memory experience. I just want those numbers above; I don't need to know the byte count, just the record count.

What I need to do to get it

 

Now, I imagine this is not implemented already for a Very Good Reason. But, enough is enough! Let's shoot for the moon and make this tool all that much better!! Anyone with me?

 

-Cedric Justice

Cambia Healthcare