Person # date tally
1 1/2/06 1
1 1/2/06 2
1 1/2/06 3
1 1/2/06 4
2 3/4/06 1
2 3/4/06 2
2 3/4/06 3
3 2/3/07 1
3 2/3/07 2
A series of readings were taken per person #. I need to create a tally from the person # column and the date column. I already have person # = to the individual readings.
I'm migrating electronic records from a legacy system and the new system has strict requirements for ASCII characters in certain metadata fields. I wrote a UDF to display illegal characters, so I can work out how to map them.
The UDF used a while loop and to improve performance, I wrote the equivalent UDF using a tally table. The tally table version actually ran significantly slower. Query calling UDF using cross apply took 26 secsfor the while loop versus 119 secs for Tally table, for test data of 97000 rows
I would like to work out why, as I will use similar code to replace the illegal characters .
-- while loop version of UDF
CREATE FUNCTION [dbo].[DisplayIllegalChars](@strText VARCHAR(4000))
RETURNS @TableVariable TABLE (
Chr CHAR(1)
,AsciiValue INT)
AS
BEGIN
DECLARE @intCount INT
DECLARE @chrCheck CHAR
[Code] .....
I have a client that I provide financial modelling services to (using Excel). They have a requirement to start capturing subscriber movements in their SQL DB. how the table should be set-up and how to extract the necessary movements report. This is largely so that I may include these components in some of the financial models that I am working on.
Subscribers are reported as follows:
Opening subs (the prior periods closing balance; or the sum of new sales at point of 1st entry)
+ New Sales (new subs)
+ Upgrades (movement from a lower product package to the associated package)
- Downgrades (movement to a lower product package from the associated package)
- Churn (subscriber losses)
Closing Balance
All transactions are captured against a specific product package, on a specific date (ymd), and for an associated platform (e.g. digital TV, broadband TV, cable TV).
I believe we only need to capture new sales, upgrades, downgrades and churn. And then used a SP to compile the movements behaviour as described above.
So perhaps the table would appear as follows:
Platform
Package
Date
Movement
Value
DTV
PROD 1
2014-11-02
New Sales
8
DTV
PROD 1
2014-11-02
Upgrades
1
[code]....
So I am assuming that given a table such as the above, we could write a SP to produce an output such as (note, below looks at monthly total so will not agree back to sample above which contains only 2 days):
Platform
Package
Movement
September
October
November
DTV
PROD 1
OPEN
600
676
776
DTV
PROD 1
New Sales
92
106
88
[code]....
how one is best to accumulate the balances given that the open date for any given reporting period is in fact an accumulation of all balances since day 1.
How does one typically capture this type of thing in SQL?
I have a table that has the following structure:
EntryID int,
Categories varchar(200)
values look like:
541,'A,B,C'
345,'B,C'
234,'A,C'
657,'D,E'
435,'D'
what I want to do is extract the Categories column to a normalized separate table:
541,'A'
541,'B'
541,'C'
345,'B' ....
I found the split using the tally table useful to split one-by-one, but how can it be applied when you are referring to a table?
I have to create a table like this across a bunch of servers. I'm thinking that I'm overlooking something with needing two additional CTEs, but maybe not. I have it at 17 seconds, which isn't much faster than a while loop solution that's currently in place.
DECLARE @START DATETIME,
@msg NVARCHAR(MAX) = N''
USE tempdb
SELECT @START = GETDATE()
CREATE TABLE dbo.EulerSource ( [SID] INT, Euler BIGINT )
[Code] ....