• Robert M. Münch (15/15) Sep 22 2002 Hi, I really like the associative arrays of D. IMO a very useable additi...
• Walter (10/18) Sep 25 2002 addition.
• Mark Evans (7/14) Sep 29 2002 Maintain an index in memory, and use that to access disk locations on de...

"Robert M. Münch" <robert.muench robertmuench.de> writes:

Hi, I really like the associative arrays of D. IMO a very useable addition.
I wouldl like to use this method for database applications. For this I have
some questions:

1. Does anybody know how to serialize such an array to harddisk, so that I
can later load it again? Any ideas how this could be done?

2. As long as I can have everything in memory using the associative arrays
is not problem. But how could a system look like that needs to read, update,
delete data from a file?

Any information is welcome.

--
Robert M. Münch
IT & Management Freelancer
Mobile: +49 (0)177 2452 802
Fax   : +49 (0)721 8408 9112
Web   : http://www.robertmuench.de

"Walter" <walter digitalmars.com> writes:

"Robert M. Münch" <robert.muench robertmuench.de> wrote in message
news:amkmhq$1vod$1 digitaldaemon.com...
 Hi, I really like the associative arrays of D. IMO a very useable

addition.
 I wouldl like to use this method for database applications. For this I

have
 some questions:

 1. Does anybody know how to serialize such an array to harddisk, so that I
 can later load it again? Any ideas how this could be done?

The only way at the moment would be to convert it to an array and then loop
through the array, see the wc.d example for how to do this.

 2. As long as I can have everything in memory using the associative arrays
 is not problem. But how could a system look like that needs to read,

update,
 delete data from a file?

The current design isn't set up for that, and it would be a hard problem to
transform it to. The best way at the moment is just serialize it as an array
for transfer to/from the disk.

Mark Evans <Mark_member pathlink.com> writes:

Maintain an index in memory, and use that to access disk locations on demand.
The memory footprint of the index will be much smaller than the actual data.
All you need as entries are begin/end file offsets associated with the data on
disk.  The index can be implemented as an associative array.

If you are into really heavy-duty requirements then use Postgresql and interface
it to D.

Mark


 2. As long as I can have everything in memory using the associative arrays
 is not problem. But how could a system look like that needs to read,

update,
 delete data from a file?

The current design isn't set up for that, and it would be a hard problem to
transform it to. The best way at the moment is just serialize it as an array
for transfer to/from the disk.