I’ve just noticed that when importing a CSV file exported from my TSB online banking, the order of transactions is inverted for transactions that occur on the same day.
For example, my online banking shows some transactions like this:
I.e the order of transactions on 10/02/2019 are inverted versus the order on the TSB statement. This isn’t a major problem, but it means it’s kind of hard to compare the TSB Online Banking statement to the quickfile statement side-by-side because the order is different.
Looking at the contents of the CSV file, they appear to be in descending date order:
A solution might be for the TSB import process to parse the CSV file into records, and then reverse the list before importing. I could obviously reverse the order of the lines in the CSV myself and that would probably fix the issue as well, but that’s a bit of a faff every time I upload a CSV file :-).
We will just insert the transactions in the way the banks output them on their CSV files.
To be completely honest there are many quirks like this with the way banks output data. It’s very difficult to write customised rules here and there to accommodate these cases. It also bloats everything on our side and we can find next week TSB have reverted the order again.
For now our priority is to focus our efforts on bank feeds.
@glenn - No worries. I can work around it - just flagging it in case it was something you didn’t know about.
@lurch - There must be some sort of secondary internal sort field on the data that remembers the order the transactions were ingested (e.g. an incrementing ID field). If I sort the statement in QuickFile by “date descending” the order of the 10/02/2019 transactions reverses, so there’s definitely a deterministic order for transactions on the same day. As Glenn says, I think this is just the order the transactions were created (or ingested from the CSV file), which means that in the case of TSB transactions the QuickFile on-screen order is the reverse of the TSB on-screen order.
From memory I think this is the case. The transactions are inserted from the CSV and assigned an incremented ID. When you’re viewing the statement it will first order by date and then by the ID field.