Hi,
I have a dataflow pulling some information from Oracle and cleaning it up - one of these steps semi-intelligently removes duplicate rows through a few bits of logic, and then blindly removes any remaining duplicates with Table.Distinct. The problem I'm getting is that although in the preview this seems to work fine, in Desktop I'm getting an error in my one-many join thanks to a duplicate record. My first instinct was that there was a case issue (e.g. 'a123' vs 'A123') as I've had these happen before, but the situation involves an index with the value always in caps.
Digging deeper, there are two columns which in the preview show a single value each, e.g:
| Index | Column 1 | Column 2 |
| FREIGHT | Value 1 | Value 2 |
Perplexingly when I download the output I have this:
| Index | Column 1 | Column 2 |
| FREIGHT | Value 1 | Value 2 |
| FREIGHT | Value 2 | Value 1 |
I'm assuming that in the original data source the latter is what is actually present in the database, this isn't being invented by Power BI, but I'm confused as to why my remove duplicates step (which is removing duplicates in the index only) does not catch this.
Any ideas?
Connor
