Ignore order and case with CSV Reconcile
The second export has the columns in reverse order, lowercase headers and its rows in a different order, with SHIPPED written as Shipped. With ignore case on, only one order, SO-302, has a real change: the total moved from 48.50 to 51.00.
- Load both files. The headers may differ in case and the columns may be in any order.
- Select the key column; Order ID and order id are treated as the same header.
- Under Comparison rules keep Ignore upper and lower case on, which also applies to the key values.
- Press Compare. Only cells that differ after trimming and case folding are listed.
- Switch Ignore case off to see how many cells would have counted as changes.
What to know
Different tools export the same table differently: columns come out in another order, headers are Title Case in one and lowercase in the other, a status is SHIPPED or Shipped or shipped. None of this is a change in the data. Columns here are matched by their names, ignoring case, so the position of a column is irrelevant, and Order ID in the first file pairs with order id in the second.
Cell values are a separate question, controlled by rules. Ignore case treats SHIPPED and Shipped as equal, and it applies to the key too: SO-304 in the first file is matched to so-304 in the second. Trimming whitespace is on by default and also collapses runs of spaces, so a trailing space from a copy and paste does not count. In the sample four orders are compared; three are unchanged despite the reordering, the case differences and the swapped columns, and SO-302 differs in the total from 48.50 to 51.00, a difference of 2.50.
Case folding can hide real differences, which is why it is a choice. Two product codes that differ only in case may be different items in some systems, and a case-sensitive password or token column should never be case-folded. Turn the rule off if case carries meaning, or ignore such columns entirely. The same applies to trimming: it is the right default for people-typed text and wrong for fixed-width codes that use padding.
Updated · Cosmovex
Questions
Does the order of columns matter?
No. Columns are matched by header name, ignoring case, so the same column in a different position is compared with itself.
Does ignore case also apply to the key?
Yes. With it on, SO-304 and so-304 are the same key. With it off, they are different keys, so one row is reported as removed and another as added.
Is whitespace compared?
By default leading and trailing spaces are trimmed and repeated spaces collapse to one before comparing. Turn that off under Comparison rules if the exact spacing matters.
Why is a row still marked as changed?
Open Changed to see the cell: it shows old -> new for each difference that survived the rules. A different number, an extra character or a different date format is the usual cause.
Can I ignore a column completely?
Yes, under Ignore columns. Ignored columns are not compared and are listed on the Columns tab.
The free plan covers everything on this page. CSV Reconcile Pro ($29, paid once) is described on the CSV Reconcile page.
