Hi All,

I'm getting a full table error on my paradox 4 table. It has 380,000
records (134MB), surely paradox handles files bigger than this. I need
about 2 million records. Does anyone known if I have hit a limit or do I re
pack the table.

Cheers

Tox
---------------------------------------------------------------------
Two men walk into a bar one after the other, you would have thought
that the second guy would have been smarter.
---------------------------------------------------------------------
This item was brought to you by the letters E and B and the number 7

2.
With 2 million records, do consider a C/S solution.

Aage J.

Aage Johansen <aagjo...@online.no> wrote in article
<32692AFC.6...@online.no>...

--

Brad
---------------------------------------------------------------------
Drink grape juice because O.J. kills
---------------------------------------------------------------------

In Pdox/Dos up-ing the block size only took effect when the table was
first created. I would assume this applies also via IDAPI. If it does
you'll need to make a new table after selecting maximum block size and
add your records to the new table.

Aage J.

The limit comes from the data-structure Paradox uses to indicate whether a
particular block is in-use or free; it places an upper limit on the number of
blocks in a single table but not upon their size.  A single block contains
many records and therefore making the block-size larger makes the capacity
greater.  It is unfortunate that BDE does not yet support specifying the
block-size at table create or restructure time.

/mr/

When loading data into a table, whether it be C/S or local, here are a few
tips and tricks to consider:

(1)  Import the records in batches.  If you need to load 100,000 records then
do them 1,000 records at a time and repeat 100 times.  (I'm talking about a
simple "for" loop in a script, of course.)  The process is therefore more
easily restartable and, generally speaking, the amount of system resources
tied up at a given point in time will be less.

(2)  If you can wrap each batch in a "transaction" then do so.  If you can get
exclusive access to the entire table during each import-batch then do so.  
This will reduce the need for locking overhead and also allow the DBMS to
cache more information in memory.

(3)  Drop all indexes and primary-keys before you begin and add them back
later.  An index can be built in a single pass through the table, but if the
index exists during the update the DBMS will try to maintain it with each and
every record that is inserted.

(4)  Back-up the table files immediately after the import is complete!

(5)  In Paradox, FLIMPORT is a doggy-dog.  Use a third-party product such as
Fast File Import or Kallista.

/mr/

Andrew
--
  Andrew Bennett
  abenn...@hyperreal.com, abenn...@cruzio.com
  http://taz.hyperreal.com/~abennett/

If you are importing this many records, you might want to remove any
indexes from the database receiving the data, and reindex after
importing. I found both pdos and interbase much faster if you do this in
two steps like this.