William,

What version?



On 20 Aug 2002, at 8:24, William Mason wrote:

> This is a multi-part message in MIME format.
> 
> ------=_NextPart_000_001A_01C24823.1181CE10
> Content-Type: text/plain;
>       charset="iso-8859-1"
> Content-Transfer-Encoding: quoted-printable
> 
> Just a follow up to the corrupted text indexes - we have been bitten =
> again by this bug.  I have never in my years of experience with R:Base =
> had so many index problems, until (it seems) that we began to use long =
> text values for keys (Text(36)) - perhaps R:Base is having trouble with =
> these.  FWIW the problem SEEMS to go away when I remove any FK =
> constraints from the text column(s).  Not a good solution, but a =
> workaround at least.  The problem only seems to occur on tables that =
> have high concurrent activity - many users pushing INSERTS at once with =
> FK values (text).  Diagnosed by comparing the number of records returned =
> by a LIST TABLE with a SELECT COUNT(*) FROM.  I've been able to sit and =
> watch the two numbers get farther and farther apart.  Users begin to =
> complain about missing or "ghost" records.  And once it starts, there is =
> no way to correct it without having all users exit and doing a reload =
> (PACK INDEX does nothing).  Again, pulling the FK's off of the table =
> seems to help.
> 
> William Mason
> 
> ------=_NextPart_000_001A_01C24823.1181CE10
> Content-Type: text/html;
>       charset="iso-8859-1"
> Content-Transfer-Encoding: quoted-printable
> 
> <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
> <HTML><HEAD>
> <META http-equiv=3DContent-Type content=3D"text/html; =
> charset=3Diso-8859-1">
> <META content=3D"MSHTML 6.00.2600.0" name=3DGENERATOR>
> <STYLE></STYLE>
> </HEAD>
> <BODY bgColor=3D#ffffff>
> <DIV><FONT face=3DArial color=3D#000080 size=3D2>
> <DIV><FONT face=3DArial color=3D#000080 size=3D2>Just a follow up to the =
> corrupted=20
> text indexes - we have been bitten again by this bug.&nbsp; I have never =
> in my=20
> years of experience with R:Base had so many index problems, until (it =
> seems)=20
> that we began to use long text values for keys (Text(36)) - perhaps =
> R:Base is=20
> having trouble with these.&nbsp; FWIW the problem SEEMS to go away when =
> I remove=20
> any FK constraints from the text column(s).&nbsp; Not a good solution, =
> but a=20
> workaround at least.&nbsp; The problem only seems to occur on tables =
> that have=20
> high concurrent activity - many users pushing INSERTS at once with FK =
> values=20
> (text).&nbsp; Diagnosed by comparing the number of records returned by a =
> LIST=20
> TABLE with a SELECT COUNT(*) FROM.&nbsp; I've been able to sit and watch =
> the two=20
> numbers get farther and farther apart.&nbsp; Users begin to complain =
> about=20
> missing or "ghost" records.&nbsp; And once it starts, there is no way to =
> correct=20
> it without having all users exit and doing a reload (PACK INDEX does=20
> nothing).&nbsp; Again, pulling the FK's off of the table seems to=20
> help.</FONT></DIV>
> <DIV><FONT face=3DArial color=3D#000080 size=3D2></FONT>&nbsp;</DIV>
> <DIV><FONT face=3DArial color=3D#000080 size=3D2>William=20
> Mason</FONT></DIV></FONT></DIV></BODY></HTML>
> 
> ------=_NextPart_000_001A_01C24823.1181CE10--
> 
> ================================================
> TO SEE MESSAGE POSTING GUIDELINES:
> Send a plain text email to [EMAIL PROTECTED]
> In the message body, put just two words: INTRO rbase-l
> ================================================
> TO UNSUBSCRIBE: send a plain text email to [EMAIL PROTECTED]
> In the message body, put just two words: UNSUBSCRIBE rbase-l
> ================================================
> TO SEARCH ARCHIVES:
> http://www.mail-archive.com/rbase-l%40sonetmail.com/
> 


================================================
TO SEE MESSAGE POSTING GUIDELINES:
Send a plain text email to [EMAIL PROTECTED]
In the message body, put just two words: INTRO rbase-l
================================================
TO UNSUBSCRIBE: send a plain text email to [EMAIL PROTECTED]
In the message body, put just two words: UNSUBSCRIBE rbase-l
================================================
TO SEARCH ARCHIVES:
http://www.mail-archive.com/rbase-l%40sonetmail.com/

Reply via email to