How to validate a form?

[homerun4711]

Thanks for your comment. I have not posted all code because the other checks that are performed are similar to the two types that I mentioned above.

Some more information:
The form consists of
- editable QLineEdits (easy to verify)
- a QTableView containing a QStandardItemModel called "tempinvoiceentries" (see code)
- some non-editable QLineEdits that are filled with customer data after a customer is selected from a popup with a QTableView containing a QSqlTableModel with the customer data (after the call the variable CustomerKey is set, see code)

I would change your return value too, returning 0 (FALSE) for OK seems a little strange.

I am confused...I thought all applications return 0 on success, why should this be different while using functions?

Thanks for the hint to use

CREATE PROCEDURE isInvoiceEntriesPresent

CREATE PROCEDURE isInvoiceEntriesPresent

To copy to clipboard, switch view to plain text mode 

This is indeed a good possibility to improve the code and will make things easier, I was not aware of this. Also thanks for the hint to create an enum, so much to learn

What exactly is the difference between COUNT(*) and COUNT(1) ?

I read the sqlite docs on this

count(X)
The count(X) function returns a count of the number of times that X is not NULL in a group.

but I don't unterstand what this exactly means.

[squidge]

Applications return an error code in which 0 is 'The operation completed successfully' and other values are pre-defined error codes. If your function only returns SUCCESS or FAILURE (like the code snippet), then the closest match to these would be TRUE and FALSE. You could typedef these values as follows:

typedef enum {FAILURE = 0, SUCCESS  = !FAILURE} ErrorStatus;

typedef enum {FAILURE = 0, SUCCESS  = !FAILURE} ErrorStatus;

To copy to clipboard, switch view to plain text mode 

COUNT(*) will retrieve all rows matching your condition (in your case, all rows, as you have no limits placed, so it will become slower over time and not scale well). It will then validate that the rows actually contain data. I'm sure when you created your schema that some fields you used the "NOT NULL" restriction to ensure they had data.

COUNT(1) doesn't care if the rows contain data or not, only that they exist.

Since your code only requires that at least one row exists, you could take it further:

SELECT COUNT(1) FROM tempinvoiceentries LIMIT 1

SELECT COUNT(1) FROM tempinvoiceentries LIMIT 1

To copy to clipboard, switch view to plain text mode 

This will always take the same amount of time to execute regardless of how many entries are in your table.

[homerun4711]

Thanks for the good information, I will rewrite some stuff now. I have a feeling that CREATE PROCEDURE will become a good friend of mine