Moving data from 5 to 8

[Orlando]

Apologies if this has already been discussed, I have been unable to find an answer for this in my search.

I am in the process of moving our current system from FileMaker 5 to 8 and because of the amount of data in some files, 50+ files with as much as 300,000 records in some files, I will be converting the files with no data. However my concern now is getting the data from the v5 files into the new v8 files in the correct fields and in the quickest way possible, as I will only have a weekend to changeover the system.

What is the best way to go about this?

Thanks for your help.

Orlando

[Oldfogey]

I've done this a few times and I think I've made every mistake you can. How are you testing your converted database without data?

Let's call the new empty Database NewA.

If you are happy with NewA and you've made any necessary changes, then -

1. Convert the old database, with data, to V8. Let's call it NewB.

2. In every table/file in NewB, write a script to Find All Records.

3. Write a single script in your main file,to execute all those individual scripts.

4. In every table/file in NewA, write a script to import records from the corresponding file/table in NewB.

5. Write a single script in your main file, to execute all those individual scripts.

6. Fix any auto-entered serial numbers you might have. It might be worth scripting that too, if you have lots of them.

The alternative is to do exactly the same thing, without scripts, for every single file.

(Do not even think about exporting from the original to an intermediate file and then importing that. I did this once!)

The benefit of scripts is that, when you mess it up, you can fix it and re-do it relatively easily.

Note that this calls for three weekends. One to make sure it all works. Second to run it. Third to fix the new problems. (Be VERY nice to your users.)

[gdurniak]

FileMaker really goofed, by not allowing imports from previous versions

I exported all data (date, time, text & number) in filemaker format, then upgraded these exports to FM8. Somewhat tedious, but it works. Import into the final copy by matching field names

[Ender]

I found the scripting more work than the actual find/import/reset serial number steps that Oldfogey described--so when I did this for my 60 file solution, I did it all manually. With the new clones ready, it took about 10 hours to convert the data, run the imports, reset serials, and test. In my case, it saved some time to run conversion and imports on a couple computers simultaneously.

Although Oldfogey doesn't like the idea of exporting to an intermediate file, if you have old files that might have had latent corruption, exporting to text first is cleaner than a direct import. I did this with a few files I had had trouble with.

One suggestion that should save time is to leave the names of fields in the new files the same as in the old files so that you can have the Import order match by name. You can rename fields later.

[Orlando]

Thanks for your responses.

The reason I am converting the files with no data is because I have read in various places that that is the best way to go about it, less corruption in the conversion. Is this still the case in FileMaker 8?

Gdurniak’s method is probably the only way to get the data with field names for matching purposes I think. Another possibility that I was considering is using ODBC, has anyone else tried this? Would this remove any latent corruption, or simply move it across?

[gdurniak]

"I am converting the files with no data is because I have read that is the best way to go about it, less corruption in the conversion"

Again, FileMaker goofed, by not providing a way to "clean", or verify, the data

If the current solution is stable, it should convert OK either way

I have posted corruption tips & info here:

http://masdevelopment.com:3455/1/57?view=print

[Ender]

The reason you might originally convert your files without data is to save time, since in that initial conversion process it might take several attempts to clean up issues in the pre-converted files. I have not heard of frequent problems with corruption during conversion. If a file fails the conversion, it is likely to fail with or without data.

In any case, I think it's easier to convert the active files to get their data rather than export to text (or .mer), then import into the new solution. Use the export/import option with any files that give you trouble.

[Oldfogey]

Ender, you are right about scripting beng more work than just doing it manually - unless you make a blunder or two.

This whole exercise is a logistics problem not a technical one. Why you would add 50+ extra steps - exportng to intermediate files - is beyond me. This creates

2*(50+) more places where Murphy can get in.

Good thought on the field names. I just took it for granted. I should have thought to mention it, because half way though one of my own upgrades, I decided to rename a field. Ouch!!

[Ender]

This whole exercise is a logistics problem not a technical one. Why you would add 50+ extra steps - exportng to intermediate files - is beyond me.

I agree on both points.