Sort exported SQL data according to dependency order

In Export Database, would it be possible to export the SQL data according to dependecy order (parent records first, then child records)? It seems to me the order is alphabetical according to table name. Output Format: SQL Object Types: Tables Generate CREATE Statements: unchecked Generate DROP Statemetns: unchecked, disabled Include Table Indexes: Unchecked Include Table Data: Checked

Hi Geff, We have received this request before so I'll add your vote for it. Finding dependencies may not always be possible, so would allowing manual ordering be a workable solution for you? Best Regards, Hans

Couldn't the map / chart in the References tab be used to manage / sort the dependencies / FKs (ignoring self-references, since that's another issue), at least for the sequence of tables to be inserted? You could add a checkbox to enable the sorted option, so that those who are OK with the current export still get it (in a much quicker way). Your other suggestion to manually order the tables is also OK, although that would be quite tiresome and still error-prone (if you don't know all the dependencies).

Hi Geff, It may be possible to use the same info as for the References tab, but it could be a slow process. There are also some details to deal with, such as if you mark that you want to export a table that has an FK to another, but you do not mark the referenced table, should it still be exported? Things like this can be worked out with various options, but it may end up being a hard-to-understand feature. Also, I'm not sure I really understand why this is an issue. The generated CREATE statements do not include any FK constraints; they are added at the end using ALTER statements. This should mean that all tables should be created without errors, regardless of dependencies. What am I missing/ Best Regards, Hans

I agree with Hans. My first thought when I read this post was "what's the use case?". My second was "Boy! That would be REALLY hard to program". Self-referencing tables would be just one issue. How do you distinguish parent/child relationships from links to lookup tables? What about circular references? My suggestion would be to name your DB tables to show the parent/child relationship. For example, a healthcare DB might have tables like Patient, Patient_Address, Patient_Insurance, etc. Then they'll sort the way you want, even in the DbVisualizer objects tree. Abbreviations can be used to make the table names shorter, as long as you establish a naming convention and stick to it.