6 replies to this topic

[TalentedFool]

Does this work?

i'm starting to design a database in which i want to use the identity type columns in SQL Server but when i look at the generated code for the SQL insert statement from Proiv It always fills in every field.

Anyway to turn that off or get Pro-iv v6 to recognise ID type columns and not generate that code?

Thanks

[Joseph Bove]

Talanted Fool,

The following may work - but I'm not sure if it's a worthwhile solution...

Create another version of the table defintion within ProIV that does not contain the identity column and use it for your writes...

hth,

Joseph

[TalentedFool]

Joseph,

Thanks for that - reading through the forum that seemed to be the way in v5 and below, was hoping that Pro-IV had done something extra in v6 to cater for identity columns.

Not really what I want to do though since it will mean I will have a whole load of dummy tables.

If I do come up with something though will let people know.

Thanks

Lee

[Rob_Stebbens]

Joseph,

Thanks for that - reading through the forum that seemed to be the way in v5 and below, was hoping that Pro-IV had done something extra in v6 to cater for identity columns.

Not really what I want to do though since it will mean I will have a whole load of dummy tables.

If I do come up with something though will let people know.

Thanks

Lee

TF.

Have you tried the "null permitted" setting in the File definition and using SETNULL() on before write?

R.

[TalentedFool]

Have you tried the "null permitted" setting in the File definition and using SETNULL() on before write?

R.

Rob

No - will see if I can give it a shot later - might not be able to do that though as some of my screens are simply paging screens which when you do a trace and see the code Pro-IV is spitting out the generated SQL includes the key field.

[Richard Bassett]

I doubt using a NULL-permissive filevar and setting it would make ProIV omit the corresponding column from a SQL INSERT.

ProIV is likely to generate 'reusable' SQL for the general case and it is very unlikely to analyze a function sufficiently to determine that a filevar is never set to anything but NULL.

OTOH, making a filevar NULL-permissive and never referring to it in your function might in principle allow ProIV to optimize the column out of an INSERT (ProIV does know which filevars are unused). I've never tested such a thing myself, let us know what you find if you try it

[Richard Bassett]

Actually, maybe this depends on what the 'semantics' of ProIV are supposed to be for adding a record.
If you don't refer to a filevar at all, do you believe ProIV should guarantee it inserted as '' or 0 even over SQL ?

If so it cannot omit the column from the SQL because there might be a SQL DEFAULT value defined.
Which I guess begs the question whether SQL DEFAULTS have a similar problem to your ID columns

Wonder what happens with DATE columns?