In the roadmap the “compare tables” feature is mentioned ? How will this feature work ?
Could this be used for validation purposes ? E.g. we have a result of an ETL process that should be compared with an already validated dataset. What we do today is to compare these datasets in EasyMorph by unpivoting these datasets to long format and then joining them. Then we do a row by row comparison.
Is the announced feature something like that ?
The “Compare tables” action will perform a series of equality checks and return a list of checks that failed (meaning the datasets are not the same). It can be used to reliably detect mismatching, however, in the initial version it will not return the exact mismatching values or their locations in datasets (only the name of the affected column will be shown). The action will detect changes such as:
- Added/removed columns
- Changed number of rows
- Column values stopped being unique
- Error values appeared/disappeared
- Empty cells appeared/disappeared
- Column has more/less text/number values
- The order of rows has changed
The new "Compare tables" action is now available in v5.6.
The action also returns the first mismatching value which should be sufficient in the vast majority of cases.