FiveTech Support Forums

Hi everyone,
I have never used CDX indexes, and I want to use them now...

How good will it be to use the TEMPORARY ( or MEMORY ) clause, on this indexes, since some say that there are problems with index file corruption some times???

Will this clause prevent this problems???

I am sure It may give me more speed if they are run in memory... Are they???

I made some test in the past with TEMPORARY clause without noticed any speed differences.

EMG

Read this
TEMPORARY
If this option is specified, a temporary index is created which is automatically destroyed when the index is closed. The temporary index may be created in memory only or in a temporary file. This lies in the responsibility of the RDD used for index creation.

My test in a LAN : non differences in elap time


exactly the same time!
bye

Marco

have you tried

INDEX ON ref TAG REF TO TEMP TEMPORARY ?

Maurizio

Maurizio wrote:Marco

have you tried

INDEX ON ref TAG REF TO TEMP TEMPORARY ?



Maurizio

Do you mean to remove FIELD from the key? I don't think it will make any difference.

EMG

Enrico

NO , change Marco with TEMP

INDEX ON field->ref TAG REF TO TEMP TEMPORARY ?

Maurizio?
This test?

ANNOUNCE RDDSYS

FUNCTION MAIN

USE i_201003

GO TOP
? seconds()
INDEX ON ref TAG REF TO TEMP TEMPORARY
? seconds()

GO TOP
? seconds()
INDEX ON field->ref TAG REF TO marco
? seconds()

RETURN NIL

INIT PROCEDURE RddInit
REQUEST DBFFPT
REQUEST DBFCDX
rddSetDefault( "DBFCDX" )

Marco

I remember reading that :

GO TOP
? seconds()
INDEX ON field->ref TAG REF TO TEMP TEMPORARY
? seconds()

This writes the file in PC memory

---

GO TOP
? seconds()
INDEX ON field->ref TAG REF TO marco TEMPORARY
? seconds()

This writes the file to hard disk

Maurizio

TEMP or marco for the index filename does not make any difference. The TEMPORARY clause only makes the difference.

EMG

I can imagine not making any difference in time, and even making everything a little slower, because of having to create indexes every time, but,

Can it make any difference in data consistency, and prevent data corruption????

Probably yes, but keep in mind that they are temporary indexes so you can't use them for the normal persistent indexes.

EMG

I’ve found an improvement using INDEX TEMP FOR … (xbrowse). Instead of SET FILTER... in LAN

Stefano

Bayron

I have never used the TEMP option .. I kinda like to know where my files are being created, their names either .idx or .cdx.

Please consider this code ..



Note that if you use .cdx the name of the file is created for you ( same name as the .dbf) with a .cdx extention. If you use .Idx .. you have to create a separate index name for each index .. .cdx is the way to go because you can build multiple ( tags ) indexes into the same .cdx.

Rick Lipkin

TEMP is not an option, it's a filename.

EMG

Thanks all for your comments...

I think that as Enrico said, Temporary indexes are good, but not for all situations...

My program does not use quite extensive Indexed databases... so I think I will go with creating permanent Indexes once per day, per week or so...

This kind of threads help us all to understand some attributes of FiveWin//(x)Harbour...

Thanks again...