January 16, 2013 at 8:52 pm
Hello Everyone
I have a rather odd query that I am trying to figure out the best way to code. I am going to try and explain the best way possible. I want to be able to do this without incorporating a loop.
I have a table that stores patient information about Diagnosis. There are three distinct hospital locations. Lets Call them LA, NY and Indy. There are three CodeNumbers for each Location : 30.1, 40.1 and 50.1
Now, what I am after is the count for each CodeNumber, per Each Location per each day.
These are the counts of each code per each location per each day.
Counts:
LA = 402 = Total
LA - 30.1 = 57
LA - 40.1 = 278
LA - 50.1 = 67
NY = 82 = Total
NY - 30.1 = 13
NY - 40.1 = 55
NY - 50.1 = 14
Indy = 142 = Total
Indy - 30,1 = 19
Indy - 40.1 = 77
Indy - 50.1 = 46
The way the data is stored is like this:
--===== If the test table already exists, drop it
IF OBJECT_ID('TempDB..#Diagnosis','U') IS NOT NULL
DROP TABLE #Diagnosis
--===== Create the test table with
CREATE TABLE #Diagnosis
(
RowID INT IDENTITY(1,1) PRIMARY KEY CLUSTERED
, DateDataRepresents Date
, CodeNumber DECIMAL(4,1)
, Location varchar(5)
)
INSERT INTO #Diagnosis
(DateDataRepresents, CodeNumber, Location)
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'NY' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'NY'
SELECT
DateDataRepresents
, CodeNumber
, Location
FROM #Diagnosis;
DROP TABLE #Diagnosis;
etc.........
The table has approx half a million rows
What is the most efficient way to get the counts?
Thank You in advance for your help, suggestions and advice
Andrew SQLDBA
January 17, 2013 at 1:53 am
Something like the following?
--===== If the test table already exists, drop it
IF OBJECT_ID('TempDB..#Diagnosis','U') IS NOT NULL
DROP TABLE #Diagnosis
--===== Create the test table with
CREATE TABLE #Diagnosis
(
RowID INT IDENTITY(1,1) PRIMARY KEY CLUSTERED
, DateDataRepresents Date
, CodeNumber DECIMAL(4,1)
, Location varchar(5)
)
INSERT INTO #Diagnosis
(DateDataRepresents, CodeNumber, Location)
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'NY' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'NY'
;with cte as
(
SELECT
DateDataRepresents
, Location
, CodeNumber
,COUNT(RowID) AS Count
FROM #Diagnosis
GROUP BY DateDataRepresents, CodeNumber,Location
with cube
)
select
DateDataRepresents,
Location,
ISNULL(CONVERT(VARCHAR,CodeNumber),'Total') AS CodeNumber,
Count
from cte where datedatarepresents is not null and location is not null order by Location,CodeNumber,DateDataRepresents
DROP TABLE #Diagnosis;
January 17, 2013 at 4:47 am
Anthony
Thanks is perfect.
Thank You. I greatly appreciate that. I never thought to use cube.
Andrew SQLDBA
January 17, 2013 at 4:49 am
If also not familiar lookup WITH ROLLUP, its another one to look at along with WITH CUBE, both produce different outputs so need to ensure you choose the right one, ROLLUP wont work in this instance but CUBE does.
January 17, 2013 at 5:06 am
Also, since this is SQL 2008, have a read up on Grouping Sets. They allow a lot more flexibility than cube and rollup do.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
January 17, 2013 at 10:16 am
GilaMonster (1/17/2013)
Also, since this is SQL 2008, have a read up on Grouping Sets. They allow a lot more flexibility than cube and rollup do.
Especially since WITH CUBE and WITH ROLLUP are being deprecated.
Drew
J. Drew Allen
Business Intelligence Analyst
Philadelphia, PA
January 17, 2013 at 12:59 pm
Thank You Everyone
Ok, I have been reading up on Grouping Sets, but I still am not getting it. How would one go about writing a query to work correctly with the same data that was given and not use Cube or Rollup?
Thanks in advance
Andrew SQLDBA
January 18, 2013 at 7:53 am
AndrewSQLDBA (1/17/2013)
Thank You EveryoneOk, I have been reading up on Grouping Sets, but I still am not getting it. How would one go about writing a query to work correctly with the same data that was given and not use Cube or Rollup?
You could always "roll your own" Totals. Performs fine with the small sample data set; if your production set is much larger, you'd want to compare metrics to see if this code scales.
--===== If the test table already exists, drop it
IF OBJECT_ID('TempDB..#Diagnosis','U') IS NOT NULL
DROP TABLE #Diagnosis
--===== Create the test table with
CREATE TABLE #Diagnosis
(
RowID INT IDENTITY(1,1) PRIMARY KEY CLUSTERED
, DateDataRepresents Date
, CodeNumber DECIMAL(4,1)
, Location varchar(5)
)
INSERT INTO #Diagnosis
(DateDataRepresents, CodeNumber, Location)
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'NY' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 40.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 50.1, 'NY' UNION ALL
SELECT '2012-12-16', 30.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'Indy' UNION ALL
SELECT '2012-12-16', 40.1, 'NY' UNION ALL
SELECT '2012-12-16', 40.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'NY' UNION ALL
SELECT '2012-12-16', 50.1, 'LA' UNION ALL
SELECT '2012-12-16', 30.1, 'Indy' UNION ALL
SELECT '2012-12-16', 50.1, 'NY'
;with cte as
(
SELECT
DateDataRepresents,
Location,
CodeNumber,
COUNT(RowID) AS Count_nb
FROM
#Diagnosis
GROUP BY
DateDataRepresents,
CodeNumber,
Location
)
--Totals by CodeNumber
SELECT
DateDataRepresents,
Location,
CAST(CodeNumber as CHAR(4)) AS CodeNumber,
Count_nb
FROM
cte
WHERE
DateDataRepresents IS NOT NULL
AND Location IS NOT NULL
UNION ALL
--Totals by Location
SELECT
DateDataRepresents,
Location,
'Total' AS CodeNumber,
SUM(Count_nb) AS Count_nb
FROM
cte
WHERE
DateDataRepresents IS NOT NULL
AND Location IS NOT NULL
GROUP BY
DateDataRepresents,
Location
ORDER BY
Location,
CodeNumber,
DateDataRepresents
DROP TABLE #Diagnosis;
Viewing 8 posts - 1 through 7 (of 7 total)
You must be logged in to reply to this topic. Login to reply