art@Elbereth:~/GoogleDrive/MyConsulting/test$ myschema -d art -t extents --myexport-express --myexport-delimiter >/dev/null art@Elbereth:~/GoogleDrive/MyConsulting/test$ ls -laFtotal 28 drwxrwxr-x 3 art art 4096 Nov 10 07:15 ./ drwxrwxr-x 75 art art 12288 Oct 29 06:27 ../ -rw-rw-r-- 1 art art 431 Nov 10 07:15 art.extents.myexport.sql -rw-rw-r-- 1 art art 502 Nov 10 07:15 art.extents.myimport.sqlart@Elbereth:~/GoogleDrive/MyConsulting/test$ cat art.extents.myimport.sql CREATE EXTERNAL TABLE extents_myimport ( dbsname "informix".lvarchar(10000), tabname "informix".lvarchar(10000), chunk INTEGER, offset INTEGER, size INTEGER, truth "informix".boolean ) USING ( FORMAT "DELIMITED", DATAFILES ( "disk:/home/art/GoogleDrive/MyConsulting/test/exten01930.unl" ), DELIMITER ",", EXPRESS, ESCAPE ON ); ALTER TABLE extents TYPE( RAW ); INSERT INTO extents SELECT * FROM extents_myimport; DROP TABLE extents_myimport; ALTER TABLE extents TYPE( STANDARD );You would just have to change the DATAFILES option to list the actual path of your CSV file.If you are not familiar with myschema, you can always download the latest release of my utils2_ak package which includes myschema, from my web site at: My Utilities
.BTW, you can speed up the dbload run somewhat by altering the table to RAW before the load. It won't be as fast as external tables, but may be fast enough for you.Art
As people have suggested you can use external tables.
Personally I would use High Performance Loader, but mainly cos I have it all scripted and ready to run.
Out of box thinking ......
· Add a timestamp and use the TimeSeries loader
· Write a simple UDR, nmap the file, scan and load.
Hi Marc,very nice challenge :-)
It depends on the process requirements: Single Table / Transaction needed / HDR Cluster installation / constraint checks / insert trigger / isolation level ...a.s.o
For the RAW table load the typical performance is from 200.000 records/second (ARM Jetson NVIDIA) to 450.000 records/second (a modern Intel x86 with enough cores for WE). All based on SD/Flash drives for the chunks and files.My estimation for this task is between 2 and 10 minutes. So you will feel the great speed of Informix with "wait less work" :-)Have funHenri