Hi George,

I found the reason for slowness. You will laugh: I forgot to select WorkFile and therefore I scaned the main database table with 1,2 mln. records in it. No wonder, why it was so slow :)

>>Hi everybody,
>>
>>I wrote a simple function to update our database with data in external file. The function looks like:
>>

>>  scan && External file
>>  if seek(PropID, "AssrMstr","PropID")
>>    replace APN with WorkFile.APN,... in AssrMstr
>>  else
>>    insert into AssrMstr ...
>>  endif
>>  endscan

>>
>>It works very slowly. The external file has ~60000 records. What could be wrong here and should I change algorithm to switch between aliases instead of using IN clause.
>>
>>Thanks in advance.
>
>How many INSERTs are typically done? How many index keys on the table? Anytime you add a record to a table with a compound, compact CDX, it has to update the index as well. If multiple keys are present this will also be true in instances where replaces are done if one of the indexed fields is modified.