import summary incorrect

[brian rich]

I'm in the process of updating an old Filemaker database of some 267000+ records where some of the fields are set as unstored calculations referencing external files and we need to change these into stored data fields. To make this change, I decided to import the fields which were unstored calculations into a temporary database to get a stored value, then reimport them back into the original database using a update match process after the corresponding fields in the original database had been turned into data fields. (The database does have one text field that is a unique key for each record, and there are no duplications as far as I can see.)

For various reasons, I ran the procedure several times on a copy of the relevant files I encountered two separate problems during this process; on two occasions the import process terminated 'normally' with the import summary box, but failed to import the complete dataset, on one occasion it failed about 30% of the way through, on the second a bit over 60%. I am certain that all records were shown in both files before the import started. On another occasion when I attempted the import, the import summary reported that 268715 records had been added/update without any errors, yet the total set of records in both databases is only 267369 - more changes than records!

Given such discrepancies, I am concerned that the import process appears unreliable with large datasets. Has anyone come across this behaviour before or suggest an alternative way of doing the field conversion without using an import process?

The import work was carried out using a local copy of FMP 11.0v3 Advanced on a Windows 7 Professional with a 2.9 GHz Intel Core2 Duo and 2 Gb memory.

TIA

Brian