User defined custom reports/layouts

[devices99]

A customer of mine is keen to have "open access" to the database to create ad-hoc reports via new layouts which aren't built into the system. As a developer I'm obviously not keen to allow table level access and to date have never provided such an feature in any bespoke system I've written for obvious reasons.

However they are a charity with a small budget and next year will need some new reports that they can't possibly know about now, and they can't afford my fees to keep changing the system for ad-hoc requirements.

I've thought about creating an export data set, into a FMP merge file which they can then open in FMP and create custom reports themselves (they have basic FMP skills only), and maybe some exports to Excel, but the system has quite a few tables and I think it's hard to ascertain in advance the link combinations they'd need.

How do others cope with this requirement? If they were a large company I'd just say it's not possible, but they are a small charity and I want to be flexible.

[comment]

As a developer I'm obviously not keen to allow table level access

Not sure what you mean by "table level access"; users can have layout editing privileges, without having full access.

the system has quite a few tables and I think it's hard to ascertain in advance the link combinations they'd need.

Perhaps you could give them a "control file" with TOs of the "real" tables and the core relationships between them. That way they could do whatever they want with the data, without messing up your file.

[devices99]

Thanks Comment, never really considered giving any direct layout access to the actual bespoke database file itself, but I concede you can lock down the tables and various other discrete items with the privileges.

However my solution is to provide essentially a control file as you mention, but rather than deal with the live data, I'll have an automatic export into this for the end-user to create custom layouts/reports.

Thanks for your response as ever!

[comment]

but rather than deal with the live data, I'll have an automatic export into this

What's the difference - other than more moving parts?