On Mon, Oct 21, 2019 at 11:17 AM, Tom H wrote:
My PBO d/l preceded the transfer by a few days and doing a manual binary search I can say that the NewID corresponds with that of g.io up to something between 1225 and 2500, above which there is a discrepancy of just 1. I'll have to redo a d/l and hope that nobody has been deleting stuff in the meantime.
I redid the PBO d/l and that 1-off discrepancy persisted. Narrowed it down to groups.io failing to import a message that it did not report as a failure. It was #1990 in plain text. So I had to revise my query of the PBO exported database to skip over that record:
SELECT id+1-(SELECT MIN(id) FROM group_message) AS Gio_ID, number AS YG_ID?
FROM group_message WHERE number<1990
UNION?
SELECT id-(SELECT MIN(id) FROM group_message) AS NewID, number AS OldID?
FROM group_message WHERE number>1990
;
So it's not only deletions on the YG occurring between the PBO and Gio transfer but also missed messages by one or the other. It can be very labour intensive finding the discrepancies. I really wish Gio had the option to preserve YG message numbers when importing to an empty database or to allow them to be offset by some user defined value that assured the resulting numbers did not clash with any already in the Gio archive.
Tom
