Count by two columns in SQL Server? - sql

Question

Here is an example table:
CREATE TABLE Example
(
LastName varchar(255),
FirstName varchar(255),
HomeAddress varchar(255),
City varchar(255)
);
INSERT INTO Example VALUES ('Murphy', 'James','123 Easy St', 'New York');
INSERT INTO Example VALUES ('Black', 'John','123 Easy St', 'Boston');
INSERT INTO Example VALUES ('Black', 'Amy','123 Easy St', 'Chicago');
INSERT INTO Example VALUES ('Simpson', 'Bill','123 Easy St', 'New York');
INSERT INTO Example VALUES ('Jones', 'James','123 Easy St', 'Chicago');
INSERT INTO Example VALUES ('Black', 'John','123 Easy St', 'Boston');
INSERT INTO Example VALUES ('Murhpy', 'James','123 Easy St', 'New York');
I want to be able to count by two columns, 'LastName' and 'City'. That is - I want this:
Name New York Boston Chicago
-------------------------------------
Jones 0 0 1
Black 0 2 1
Simpson 1 0 0
Murphy 2 0 0
My question is similar to this question. I created the following, which produces very close results:
DECLARE #DynamicPivotQuery AS NVARCHAR(MAX),
#PivotColumnNames AS NVARCHAR(MAX),
#PivotSelectColumnNames AS NVARCHAR(MAX)
--Get distinct values of the PIVOT Column
SELECT #PivotColumnNames = ISNULL(#PivotColumnNames + ',','') + QUOTENAME(City)
FROM (SELECT DISTINCT City FROM Example) AS cat
--Get distinct values of the PIVOT Column with isnull
SELECT #PivotSelectColumnNames
= ISNULL(#PivotSelectColumnNames + ',','')
+ 'ISNULL(' + QUOTENAME(City) + ', 0) AS '
+ QUOTENAME(City)
FROM (SELECT DISTINCT City FROM Example) AS cat
--Prepare the PIVOT query using the dynamic
SET #DynamicPivotQuery =
N'SELECT LastName, ' + #PivotSelectColumnNames + '
FROM Example
pivot(count(City) for City in (' + #PivotColumnNames + ')) as pvt';
--Execute the Dynamic Pivot Query
EXEC sp_executesql #DynamicPivotQuery
BUT, this is not quite what I am after. It produces this:
I tried to add 'GROUP BY LastName' to the end of the query, but it does not work. I get an error:
Column 'pvt.Boston' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause.

Answer 1

You are so close!
Two things:
Use a subquery/derived table to select only the columns you need for your pivot
Don't misspell Murphy
--Prepare the PIVOT query using the dynamic
SET #DynamicPivotQuery =
N'SELECT LastName, ' + #PivotSelectColumnNames + '
FROM (select LastName, City from Example) e
pivot(count(City) for City in (' + #PivotColumnNames + ')) as pvt';
rextester demo: http://rextester.com/ETJ57985
returns:
+----------+--------+---------+----------+
| LastName | Boston | Chicago | New York |
+----------+--------+---------+----------+
| Black | 2 | 1 | 0 |
| Jones | 0 | 1 | 0 |
| Murphy | 0 | 0 | 2 |
| Simpson | 0 | 0 | 1 |
+----------+--------+---------+----------+

Answer 2

I think simple pivot will give your query results
Select * from (
Select LastName, City from Example ) a
pivot (count(city) for city in ([New York],[Boston],[Chicago])) p
If you change your query as below you will get appropriate results:
--Prepare the PIVOT query using the dynamic
SET #DynamicPivotQuery =
N'Select LastName, ' + #PivotSelectColumnNames + ' from ( SELECT LastName, City
FROM Example ) a
pivot(count(City) for City in (' + #PivotColumnNames + ')) as pvt';
Output as below:
+----------+----------+--------+---------+
| LastName | New York | Boston | Chicago |
+----------+----------+--------+---------+
| Black | 0 | 2 | 1 |
| Jones | 0 | 0 | 1 |
| Murhpy | 1 | 0 | 0 |
| Murphy | 1 | 0 | 0 |
| Simpson | 1 | 0 | 0 |
+----------+----------+--------+---------+

Answer 3

Try this -
DECLARE #DynamicPivotQuery AS NVARCHAR(MAX),
#PivotColumnNames AS NVARCHAR(MAX),
#PivotSelectColumnNames AS NVARCHAR(MAX)
--Get distinct values of the PIVOT Column
SELECT #PivotColumnNames = ISNULL(#PivotColumnNames + ',','') + QUOTENAME(City)
FROM (SELECT DISTINCT City FROM #example) AS cat
--Get distinct values of the PIVOT Column with isnull
SELECT #PivotSelectColumnNames
= ISNULL(#PivotSelectColumnNames + ',','')
+ 'ISNULL(' + QUOTENAME(City) + ', 0) AS '
+ QUOTENAME(City)
FROM (SELECT DISTINCT City FROM #example) AS cat
--Prepare the PIVOT query using the dynamic
SET #DynamicPivotQuery =
N'SELECT LastName, ' + #PivotSelectColumnNames + '
FROM #example
pivot(count(City) for City in (' + #PivotColumnNames + ')) as pvt group by LastName, ' + #PivotColumnNames;
--Execute the Dynamic Pivot Query
EXEC sp_executesql #DynamicPivotQuery

Answer 4

Alternative approach not using PIVOT:
DECLARE #sql NVARCHAR(max);
SET #sql = ''; -- not necessary for CONCAT, but necessary if converting this answer to + style string concatenation
SELECT
#sql = CONCAT(#sql, ', COUNT(CASE WHEN city = ''', City, ''' THEN 1 END) as ', QUOTENAME(City))
FROM
Example
GROUP BY City
SET #sql = CONCAT('SELECT Name', #sql, ' FROM example GROUP BY name')
EXEC sp_executesql #sql