Path: utzoo!attcan!uunet!ncrlnk!ncrcae!hubcap!gatech!rutgers!rochester!pt.cs.cmu.edu!andrew.cmu.edu!rl1b+
From: rl1b+@andrew.cmu.edu (Robert A. Locke)
Newsgroups: comp.databases
Subject: Re: dBASE III+ bug?
Message-ID: <QY6pBUy00VQFM8iGMI@andrew.cmu.edu>
Date: 13 Mar 89 05:36:32 GMT
References: <6056@bsu-cs.UUCP>,
	<21458@agate.BERKELEY.EDU>
Distribution: na
Organization: Academic Computing, Carnegie Mellon, Pittsburgh, PA
Lines: 24
In-Reply-To: <21458@agate.BERKELEY.EDU>

Paul,

You were correct in your assessment from the second message.  It does have to do
with your using STR without specifying the length.    The function creates a
varying length key which dBASE cannot handle.  Since you were doing an APPEND
BLANK, the length of the field was set to zero.  When you do an APPEND, it sets
it to a more understandable length (and I believe, takes on the length specified
in the field definition).  This is why the first one corrupts your index and the
second one does not.

Now, as to the way you should probably solve your problem:

When you specify the numeric field in your database there will be two numbers
associated with it (length & decimal places).  Simply use these numbers in the
STR(field,length,decimal) in your index key and everything will work fine.

--Rob Locke
-------
Arpanet: rl1b+@andrew.cmu.edu
Bitnet : rl1b@ANDREW
Fidonet: Rob Locke (1:129/15)
USNail : 5419 Normlee Place
         Pittsburgh, PA  15217
         USA