Plateforme Level Extreme
Abonnement
Profil corporatif
Produits & Services
Support
Légal
English
HyperFile Performance
Message
De
15/01/2013 09:41:02
 
 
À
Tous
Information générale
Forum:
WinDev
Catégorie:
Autre
Titre:
HyperFile Performance
Divers
Thread ID:
01562573
Message ID:
01562573
Vues:
87
Just a bit of database performance info.


Setup: For VFP we use a DBF/CDX. For WinDev we use a FIC/NDX (which are the file extension names WinDev uses for its built-in HyperFile Database. FIC extension is synonymous DBF and NDX is synonymous with CDX. The NDX is not the same as in xBase as it is a compound index file and, I assume, propriety to WinDev.


Test 1. Sequentially read +/-21,000 records which have a specific index key value. File is read in index order and contains +/-10m records and about 800Mb in size.

In both WinDev and VFP the pseudo code is:
seek( condition ) 
scan while condition

  field1 = some_field  &&  Just adding some read operation overhead 
  field2 = some_field

endscan
VFP - my average time after several iterations/tests = +/- 0.04 seconds
WinDev 64 bit - my average time after several iterations/tests = +/-0.3 seconds

WinDev is roughly a factor of 10 slower than VFP but still the result set of 21,000 records is certainly returned fast enough for my purposes - ymmv. Since my experiences in WinDev is still relatively new there might be a faster way of doing this in WinDev and I have sent them a request to check that my usage of WinDev functions is optimal.


Test 2. SQL SELECT a set of +/-21,000 records matching an index key field from a file containing +/- 10m records and about 800Mb in size.

In VFP you just SELECT YOUR-SQL-STATEMENT.

In WinDev you need to create the SQL Select in a string and pass it to a function to be executed against a file. You then get a result set that you can process or loop through or whatever.

VFP - my average time after several iterations/tests = +/-3 seconds
WinDev 64 bit - my average time after several iterations/tests = +/-0.4 seconds

Here WinDev is roughly a factor of 10 times faster than VFP in the SQL SELECT test. I am guessing they have spent more time on optimizing their SQL functionality.
In the End, we will remember not the words of our enemies, but the silence of our friends - Martin Luther King, Jr.
Suivant
Répondre
Fil
Voir

Click here to load this message in the networking platform