Issue 83250 - DataPilot: refactor backend data access to use cache table
Summary: DataPilot: refactor backend data access to use cache table
Status: CLOSED FIXED
Alias: None
Product: Calc
Classification: Application
Component: code (show other issues)
Version: 680m233
Hardware: All All
: P3 Trivial (vote)
Target Milestone: ---
Assignee: oc
QA Contact: issues@sc
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2007-11-02 15:11 UTC by kyoshida
Modified: 2013-08-07 15:15 UTC (History)
3 users (show)

See Also:
Issue Type: ENHANCEMENT
Latest Confirmation in: ---
Developer Difficulty: ---


Attachments
Example for the drill-down problems (10.10 KB, application/vnd.oasis.opendocument.spreadsheet)
2008-04-09 12:49 UTC, niklas.nebel
no flags Details

Note You need to log in before you can comment on or make changes to this issue.
Description kyoshida 2007-11-02 15:11:27 UTC
We will switch to using cache table to cache the original data to avoid multiple
scanning of source data, which is expensive especially for external database
data.  We will also simplify the current data access code by making effective
use of the cache table implementation.

This work is on-going in koheidatapilot01 CWS.
Comment 1 kyoshida 2007-11-02 15:11:53 UTC
accepting the issue.
Comment 2 kyoshida 2007-11-02 15:13:19 UTC
Setting target to 3.0.
Comment 3 kyoshida 2008-03-14 23:23:32 UTC
fixed in koheidatapilot01.
Comment 4 kyoshida 2008-03-14 23:25:57 UTC
re-assigning to oc for QA.
Comment 5 oc 2008-04-01 08:04:50 UTC
verified in cws
Comment 6 niklas.nebel 2008-04-09 12:47:57 UTC
The options "Ignore empty rows" and "Identify categories" no longer work. These
aren't the most important options, but they are there, so they must work.

"Drill down" on a data field is a bit different from what it was in 2.4:

- In 2.4, group columns were fully usable, so one could view the data for the
subtotal of a group (see attached file). This wasn't specified explicitly, but
it's always problematic to remove something that once worked. You should at
least check if this can be added to your implementation without too much effort.

- If a column/row field contains numeric entries, comparison should be done on
the values (and was in 2.4). Otherwise, entries with different number formats
might not be recognized, an example is also in the attached file.
Comment 7 niklas.nebel 2008-04-09 12:49:07 UTC
Created attachment 52685 [details]
Example for the drill-down problems
Comment 8 oc 2008-04-14 14:38:18 UTC
reopened because of above mentioned regression
Comment 9 oc 2008-04-14 14:39:52 UTC
back to development
Comment 10 kyoshida 2008-04-14 15:30:15 UTC
@nn: thanks for the regression report.  I just got back from Prague so I'll be
little tied up for the next few days, but I'll look into this.
Comment 11 kyoshida 2008-04-17 14:44:07 UTC
>The options "Ignore empty rows" and "Identify categories" no longer work. These
aren't the most important options, but they are there, so they must work.

This is fixed, and now should work.

>- In 2.4, group columns were fully usable, so one could view the data for the
subtotal of a group (see attached file). This wasn't specified explicitly, but
it's always problematic to remove something that once worked. You should at
least check if this can be added to your implementation without too much effort.

This was a bit of work, but it should work now.  I hope I got it right in all cases.

>- If a column/row field contains numeric entries, comparison should be done on
the values (and was in 2.4). Otherwise, entries with different number formats
might not be recognized, an example is also in the attached file.

This should also work.  I now use ScDPItemData instead of plain-old String class
to carry the data around.  Let me know if there are still cases where it fails.

All fixes are checked in to koheidatapilot01 cws.
Comment 12 kyoshida 2008-04-17 14:53:19 UTC
RESOLVED FIXED once again.
Comment 13 kyoshida 2008-04-20 07:13:53 UTC
back to oc for QA.
Comment 14 kyoshida 2008-04-21 02:54:50 UTC
forgot to reopen for QA.
Comment 15 kyoshida 2008-04-30 18:27:41 UTC
setting the status to FIXED
Comment 16 oc 2008-05-14 12:01:53 UTC
verified in internal build cws_koheidatapilot01
Comment 17 oc 2008-09-22 13:47:55 UTC
closed because fix available in builds OOO300_m6 and DEV300_m31