Hi dieter,
Some change in requirement.
Table structure has changed. reporting columns are static.
They are from Jan -15 to Dec - 17.
CREATE TABLE subscriber_feature
(SUB_ID INTEGER, -- subscriber id
feature_code VARCHAR(20) -- feature code
feature_add_dttm DATE, --add dttm
feature_remove_dttm DATE, -remove dttm
feature_cancel_dttm DATE, -- cancel dttm
Sub_stat_cd CHAR(1)) - ex value 'a' - ACTIVE, 'c' - CANCEL etc
PRIMARY INDEX(sub_id);
INSERT INTO subscriber_feature VALUES(123,'Code','2015-01-15','2015-03-20', NULL,'A' );
INSERT INTO subscriber_feature VALUES(234,'Code','2015-01-20','2015-04-05', '2015-05-02','C');
INSERT INTO subscriber_feature VALUES(345,'Code','2015-02-05','2015-04-15',NULL,'A');
INSERT INTO subscriber_feature VALUES(456,'Code','2015-01-25',NULL,2015-03-21,'C');
INSERT INTO subscriber_feature VALUES(456, 'Code','2015-04-20','2015/05/20',NULL,'A' );
INSERT INTO subscriber_feature VALUES(567, 'Code','2015-02-10','2015-02-25',NULL,'A');
INSERT INTO subscriber_feature VALUES(678, 'Code','2015-03-10',NULL,NULL,'A');
One problem in that. If the customer removed feature or cancel subscription in same month he added, then end_dt will give previous month end date.
So count may go wrong for this scenario.
We are considering 6 million records. Can we introduce some table to store intermediate result in between to make the query performance better. If so, what format it should be?
Looking forward for your help.
Thanks
You can see this link: http://www.info.teradata.com/ . In it, you can find "Teradata Database, Tools and Utilities Release". You can choose which version you are in(13.10,14,....). I suggest you start with introduction, then step by step with others like "SQL Data Definition Language Detailed Topics" so that you can see the real beauty, how it is different from other DBs.