I was able to load 100M rows into a single table from a FLAT file. This clearly shows that the following problem is not related to a SIZE issue but something else.
 
Thanks,

 
On 5/10/07, Venks <venkatesh.mantha@gmail.com> wrote:
Hi,
 
I have installed the latest version of MonetDB using the all-in-one script on a 64 Bit Linux (Redhat) with 4GB of memory. I created 3 tables containing 10Million rows each.
 
The table definition is as follows:
 
CREATE TABLE data_1mto10m (
id      INTEGER,
col1   INTEGER,
col2   VARCHAR2(5)
) ;
 
The above table definition applied to all remaining tables too. The first column "id" is the primary key which is nothing but a running serial number from 1 to 10,000,000.
 
CREATE TABLE data_10mto20m ( ...
 
CREATE TABLE data_20mto30m ( ...
 
For the second table the "id" column starts from 10,000,001 and ends with 20,000,000. For the third table the "id" column starts from 20,000,0001 and ends with 30,000,000.
 
The second and third column contains some randomly generated numbers.
 
After that I am trying to create another table to include data from all the above as follows:
 
1. CREATE TABLE data_1mto30m AS SELECT * FROM data_1mto10m WITH DATA ;
2. INSERT INTO data_1mto30m SELECT * FROM data_10mto20m ;
3. INSERT INTO data_1mto30m SELECT * FROM data_20mto30m ;
 
In shot, I am trying to create a 30 million row table by including all the data from the three 10 million row tables.
 
After I complete the above steps and close the database, I am NOT able to re-open the database again. I get the following error:
 
>mserver5: /scratch/monet.GNU.64.64.d.9712/MonetDB/src/gdk/gdk_utils.mx:1092: GDKfree: Assertion `size != 0' failed.
 
I understand that MonetDB can handle huge tables running on a 64 bit system and would like to test the same.
 
Please help me in resolving this issue.
 
Thanks,
 
- Venkatesh