External Tables

[qube99]

What are the pros and cons on using external tables? My training vids don't explain the logic for deciding.

What I have in mind is keeping the data separate from the interface. I'd like to understand the issues better before heading down this path.

[Ocean West]

well depending on the content of data you can leverage server resources, you could put data tables on under used servers to maximize efficiency. Braking up data tables based on logical breaks in your solution - modularize your tables by function or by phase of a project, or departments, or audience context, such as a module for IWP or PHP users you could have different file where users are granted access to the file.

If you are making shrinkwrap software and your schema or data tables are locked down you can distribute updates easier as the end users data elements don't change but the interface and solution logic can - just replace the interface file.

Your data tables are free of UI related fields, pure and clean data. Data tables have limited scripts and relationships as much of the logic is contained in the user interface file.

One CON is you have to replicate your custom functions and security schema across all files. (consider looking in to external authentication - if applicable)

You have to be cognizant that you will be file hopping and window juggling especially during early development FMP doesn't make it easy to toggle between the two separate files. - If your deep in one schema in one file you have to close all dialog boxes to back out switch files and drill back in.

Don't use native FileMaker value lists with the option for the user to edit them, in your interface file if you ever intend to replace the file, as you have no easy way to preserve the data the client has entered in to them.

There is a bug in FileMaker when using the separation model where you get in I believe it is relationships in scripts when targeting a related table and related layout. - It shows up in your DDR reports.

Here is a good technique too http://www.filemakermagazine.com/videos/developer-tactics.html

[qube99]

Good info. Thanks.

[MorFologist]

[color:blue]"There is a bug when using the separation model where you get ."

I wondered why I see that on occasion. Thanks for clearing up the mystery!

One other point to add - with separate data files you can decide which to backup when, and with what frequency. The huge file with the related container storage can be left to backup once in the middle of the night.