Comments on: Data Extraction into SQL Server Database from Cache ODBC Database

By: daveinaz

daveinaz — Fri, 20 Jan 2006 10:45:59 +0000

Boy, am I confused! First you say you “doubt the database providers will be willing to supply the schema”, but then you say you “have the structure of the tables in the system” and you “know is that the table is a view and not an actual table”. And your query returns different results each time you run it? Jeez!

Ok, let’s start with some questions for you.
1 Who are these db providers? Are they outside your company? Who does this cache db belong to? If it belongs to your company, can’t you negotiate a deal to get the source for the view you’re trying to emulate? And, if they let you have access to their data, why wouldn’t they be willing to share the SQL source behind the view?

2 How did you recreate the table structure when you don’t know the schema?

3 Which table “is a view, not a table”? The _SYSTEM.Patient table in your querying?

4 You say you want to “re-create the view using joins”, but your SQL only includes one table. Why?

5 Do you KNOW that the resultset is larger than it should be, or are you guessing? If you do know, HOW do you know? Are you able to run the orginal query you’re trying to reproduce?

6 If the recordcounts vary greatly, not just growing, and this db is truly a “cache database”, perhaps the cache is being flushed periodically (as it should be, if it’s really a cache). That is, maybe it’s being emptied and then refilled, but not necessarily with the same data. Caches are often used to temporarily “stage” data prior to some operation like printing invoices or reports.

Sorry for all the questions, but I don’t think we have a very clear picture of the situation, so it’s hard to know what to advise.

By: ferencmantfeld

ferencmantfeld — Wed, 18 Jan 2006 13:41:07 +0000

roughly how many records are we talkng about in the query ?

By: spiddy

spiddy — Wed, 18 Jan 2006 11:58:05 +0000

Hi all, Many thanks for your replies.
I am using DTS to extract the data using a simple select statement in a transform data task SQL Query, and using a date criteria to reduce the size of the extract like this:
SELECT AKA, Admitting_Clinician, Blood_Bank_Work_Issued, Blood_Group, ??/* Up to about 100 columns */
FROM _SYSTEM.Patient
WHERE Date_Patient_Details_Amended >= ’2005-04-01′
AND Date_Patient_Details_Amended

By: daveinaz

daveinaz — Wed, 18 Jan 2006 10:05:26 +0000

Well, I hate to disagree with the previous posters, but it sounds to me like you SHOULD be querying the view, NOT the tables. You say the view returns the “correct”, expected number of records, so clearly it already contains the necessary joins and criteria. Why duplicate that effort?

I do agree that DTS is the way to go, though. Just point it at the view and let ‘er rip.

By: upesh1

upesh1 — Tue, 17 Jan 2006 12:12:26 +0000

hi,
u should try getting data from the base by using left & right joint & check the joine. i would like to see yr query which is give better idea for yr problem.u might be use view with multiple column along with child table.
there is 1 way to get data from cache database in sqlserver is dns. by this u can extract data.