Problem creating a new record

[raptor]

Hello Experts

I have a problem creating a new record in one table. At the moment, the next sequence number should be shown, nothing happens anymore. The client is not responding, the process 'prowin32.exe' is consuming a log of processor-capacity. The problem appears only in one table. Does anybody know how to solve this problem? Where are actual sequence-values stored?

Best regards
Troubleshooter

 

[TomBascom]

You're going to need to write something a bit clearer about what is going on and what you need help with.

 

[raptor]

create table_xy.
assign field1 = "a".
assign field2 = "b".

running this, the client doesn't respond anymore. The problem appears only in one table. I think (...) that the problem is somewhere in sequences.

 

[Stefan]

Are table / field triggers defined for this table? If so, what are they doing?

Both can be easily verified by running your code with the debugger.

Alternatively / additionally you can use logging (see help on log-manager system handle) to view all statements executed.

 

[raptor]

Thank you for your answer. I will check triggers as soon as I will be back in the office. But anyway, it cannot be a problem of the triggers. The problem appears since a 2 days, before it worked for years...

 

[Stefan]

Thank you for your answer. I will check triggers as soon as I will be back in the office. But anyway, it cannot be a problem of the triggers. The problem appears since a 2 days, before it worked for years...

Triggers can contain all sorts of fun which can be data dependent - so how do you manage to rule this out this quickly?

 

[raptor]

This is the create-trigger of this table:

TRIGGER PROCEDURE FOR CREATE OF mailing.

def var l-ml_nr like mailing.ml_nr.
def buffer b-mailing for mailing.

{syvar.i}
{symon.i}

IF g-during_load = false
THEN DO:
    BLOCK:
  do while true:
    assign l-ml_nr = next-value(next-ml_nr).
    find b-mailing where 
         b-mailing.ml_nr = l-ml_nr
         no-lock no-error.
    if not available b-mailing then leave.
                               else next.
  end.
  if l-ml_nr = 0 then l-ml_nr = next-value(next-ml_nr).  
  assign mailing.ml_nr   = l-ml_nr
         mailing.sy_user = g-sb_code.
END.

The problem appears in the 'do while true' area. The line with 'if...' is executed anymore.


The Code

message next-value(next-ml_nr)

shows a value.


There are a few similar triggers for similar tables. Everything worked for years. Since yesterday this one table has this problem.

 

[Stefan]

This is the create-trigger of this table:

TRIGGER PROCEDURE FOR CREATE OF mailing.

def var l-ml_nr like mailing.ml_nr.
def buffer b-mailing for mailing.

{syvar.i}
{symon.i}

IF g-during_load = false
THEN DO:
    BLOCK:
  do while true:
    assign l-ml_nr = next-value(next-ml_nr).
    find b-mailing where 
         b-mailing.ml_nr = l-ml_nr
         no-lock no-error.
    if not available b-mailing then leave.
                               else next.
  end.
  if l-ml_nr = 0 then l-ml_nr = next-value(next-ml_nr).  
  assign mailing.ml_nr   = l-ml_nr
         mailing.sy_user = g-sb_code.
END.

The problem appears in the 'do while true' area. The line with 'if...' is executed anymore.

See... lot's of fun in the trigger.

The Code

message next-value(next-ml_nr)

shows a value.


There are a few similar triggers for similar tables. Everything worked for years. Since yesterday this one table has this problem.

And which value is it showing? Has the sequence wrapped around (we do not use sequences, so no idea what happens when a sequence hits it's maximum) resulting in the DO WHILE TRUE loop racing through all your old mailing records.

Turn on clientlogging / use the debugger to see what's happening.

 

[raptor]

Fun is something different for me.

Sequence limit can be the reason, i will check it immediately.

 

[raptor]

Ok, the problem was in the max-value of the sequence. Thank you for your time and have a nice weekend!

 

[jdpjamesp]

And which value is it showing? Has the sequence wrapped around (we do not use sequences, so no idea what happens when a sequence hits it's maximum)

Just for completeness, you can decide whether it wraps or not when you create the sequence, along with a few other options such as how it increments, where it starts and where it stops.