One week into DS30, and we've already covered a lot of ground...
The past few days have been all about introducing us to the basics of Alteryx and Tableau Prep for extracting, transforming, and loading data; it's been great to go from learning theory to completing challenges even at this early stage.
While getting familiar with the many tools in Alteryx, a few concepts challenged me in particular.
For my very first blog, I decided to record some of the common pitfalls that I will personally need to keep in mind as I move forwards.
'Join' Jargon
The different types of Join – and how these differ from Tableau – need careful attention. In Alteryx a Join’s output is broken down into ‘L’, ‘J’ and ‘R’.
- ‘L’ output gives us the ‘Left Unjoined’ data, which is any information from the first table that had no data under certain field headings in the second table.
- ‘J’ output gives us the results of the ‘Inner Join’, which is all the data from both tables that got joined together based on a field they had in common.
- ‘R’ output gives us the ‘Right Unjoined’ data, which is information from the second table that was not joined because it did not have datapoints under certain field headings in the first table.
In Tableau, however, a ‘Left Join’ includes ‘Left Unjoined’ data its output, simply leaving 'null' values in any unpopulated fields.
Keep this in mind when switching between Tableau and Alteryx.
Ending Conditional Expressions
In another departure from Tableau syntax, Alteryx uses ‘ENDIF’ to end a conditional expression, instead of just ‘END’. This has already caught me out several times…
Setting Data Type
Many issues will arise in a workflow if the wrong data types are used for each field. When working with numbers in a calculation, be sure to change data type from a string to an integer type, or the ‘Double’ or ‘Fixed Decimal’ types; this can be done using tools like ‘Select’.
When working with decimals, it is worth keeping the values to more decimal places while doing calculations, and then rounding to fewer decimal places right at the end if needed: rounding up too early in the flow could affect the accuracy of your final output.
'Append' Errors
When using the ‘Append’ tool, the automatic setting generates an error when 16 or more records are returned – remember to uncheck this option if expecting more than 16 records in the output!
And finally…
One Formula Will Do!
Remember that there is a little plus sign in the ‘Formula’ tool for adding more expressions – we can avoid creating a little string of test tube icons in a row, and be more efficient with just one tool.
I look forward to covering more ground in the weeks ahead, and (hopefully) getting more confident in each of these areas.