Errors, EOFs, and a successful query

[ids]

In a system module, during an system_update function, I am trying to avoid the "fatal error" scenario when I try to update a row and values already exist. In my query, I am trying to set an error or flag so I can report the reason an update didn't execute to the content screen so the user knows there was an input error and why.

In theory, I am testing the myview.d.eof value. If there is EOF, the query didn't find a record. If not myview.d.eof, then a record already exists and I need to report the entry error.

    <MvIF EXPR = "{ myview.d.EOF }">
        <MvCOMMENT>-SDS--no duplicate found </MvCOMMENT>
        <MvCLOSEVIEW NAME = "Merchant" VIEW = "myview">
        <MvFUNCTIONRETURN VALUE = "0">
    <MvELSE>
        <MvCOMMENT>-SDS--at least one record was found </MvCOMMENT>
        <MvCLOSEVIEW NAME = "Merchant" VIEW = "myview">
        <MvFUNCTIONRETURN VALUE = 1>
    </MvIF>

Doing the query with known values, all scenarios have the result that myview.d.EOF always fails and at least one record is found.

Why might myview.d.EOF not be true when no records are found in the query? is there another status I should be checking instead or a better way to accomplish this?

Note: I've have an Unique Index on the two fields in question, which I assume are the reason for the Fatal Error. Should I consider not having these fields as Unique Indexes?

Thanks,

Scott

[Kent Multer]

If functions such as Module_System_Update return a zero, that will always trigger a fatal error for Merchant. I've written lots of system-ex modules, and they always return 1. If some type of user error occurs during Module_System_Update, my code will set a global variable to indicate the error. That variable will still exist when the store calls Module_System_Content; so I'll put some code in there to check the variable and display an error message.

[ids]

Thanks, Kent.

I was in a hurry to get out to the office earlier and neglected to specify the query. The query I'm executing is a function executed in the system_update. The return value of 1 or 0 is strictly based on the EOF flag in that function when I do the query -- which asks, does a record already exist?

I have a couple of other ideas about how to tackle this but this:

<MvIF EXPR = "{ myview.d.EOF }">
   <MvCOMMENT>-SDS--no duplicate found </MvCOMMENT>
   <MvCLOSEVIEW NAME = "Merchant" VIEW = "myview">
   <MvFUNCTIONRETURN VALUE = "0">
<MvELSE>
   <MvCOMMENT>-SDS--at least one record was found </MvCOMMENT>
   <MvCLOSEVIEW NAME = "Merchant" VIEW = "myview">
   <MvFUNCTIONRETURN VALUE = 1>  
</MvIF>

Isn't written to find an error, it's written to gain the status. When I know the status, I can report the status but it also can trigger the table update function running in the system_update.

My basic question, why would it be that myview.d.EOF is failing to be true?

Scott

[Kent Multer]

Sorry I was off-target there. I can't tell more without seeing the rest of the code; but I can tell you that I've never found the d.eof flag to be inaccurate. I would put in some debugging messages, and/or test my SQL query using phpMyAdmin, in search of some other cause for the problem.

[ids]

I've also not ever experienced the EOF to be inaccurate either. Hence, baffled.

Scott

[Gordon - Phosphor Media]

Scott, I also have used .EOF for years and never has it failed me. If you misspelled the view, though - wouldn't .EOF on a nonexistent view always be false?

I can't readily think of anything else that would 'switch off' your experience of using .EOF.