How use count for value from another table - sql

Question

How to use count joining with different tables. Please have a look at my queries. Here i am using CROSS APPLY. But i am not getting the actual result.
how i can to get all the item from item table not in incident table.
Tabel : Inc_cat
+------------+--------------+--+
| inc_cat_id | inc_cat_n | |
+------------+--------------+--+
| 1 | Support | |
| 2 | PM | |
| 3 | Installation | |
+------------+--------------+--+
Table:incident
+-------------+---------+------------+-----------------+
| incident_id | item_id | inc_cat_id | date_logged |
+-------------+---------+------------+-----------------+
| 100 | 555 | 1 | 2016-01-01 |
| 101 | 555 | 2 | 2016-01-18 |
| 103 | 444 | 3 | 2016-02-10 |
| 104 | 444 | 2 | 2016-04-01 |
| 105 | 666 | 1 | 2016-04-09 |
| 106 | 555 | 2 | 2016-04-20 |
+-------------+---------+------------+-----------------+
Table:item
+---------+---------+--+
| item_id | cust_id | |
+---------+---------+--+
| 444 | 34 | |
| 555 | 34 | |
| 666 | 76 | |
| 333 | 34 | |
| 222 | 34 | |
| 111 | 34 | |
+---------+---------+--+
Result:
+---------+----------------+-----------+---------------------+
| item_id | count(Support) | count(PM) | count(Installation) |
+---------+----------------+-----------+---------------------+
| 555 | 0 | 1 | 0 |
| 444 | 0 | 1 | 0 |
| 666 | 0 | 0 | 0 |
| 333 | 0 | 0 | 0 |
| 222 | 0 | 0 | 0 |
| 111 | 0 | 0 | 0 |
+---------+----------------+-----------+---------------------+
My Query:
SELECT i.item_ID,
COUNT(CASE WHEN i.inc_cat_id = ic.inc_cat_id AND i.inc_cat_id = 1 THEN 1 END) AS cntSupport,
COUNT(CASE WHEN i.inc_cat_id = ic.inc_cat_id AND i.inc_cat_id = 2 THEN 1 END) AS cntPM,
COUNT(CASE WHEN i.inc_cat_id = ic.inc_cat_id AND i.inc_cat_id = 3 THEN 1 END) AS cntInstallation
FROM #incident i
CROSS APPLY #incCat ic
WHERE (i.date_logged BETWEEN '2016-04-01' AND '2016-04-30')AND i.cust_id='34'
GROUP BY i.item_ID

Answer 1

You don't need a CROSS APPLY. A simple LEFT JOIN will do:
SELECT i.item_id,
COUNT(CASE WHEN inc.inc_cat_id = 1 THEN 1 END) AS cntSupport,
COUNT(CASE WHEN inc.inc_cat_id = 2 THEN 1 END) AS cntPM,
COUNT(CASE WHEN inc.inc_cat_id = 3 THEN 1 END) AS cntInstallation
FROM Item AS i
LEFT JOIN Incident AS inc ON i.item_id = inc.item_id AND
inc.date_logged BETWEEN '2016-04-01' AND '2016-04-30'
WHERE i.cust_id = 34
GROUP BY i.item_id
You just need to start by table Item, so as to get all items returned, as in the expected result set in the OP.
Demo here

Answer 2

Unless I'm missing something, your query is over complicated:
SELECT item_ID,
COUNT(CASE WHEN .inc_cat_id = 1 THEN 1 END) AS cntSupport,
COUNT(CASE WHEN inc_cat_id = 2 THEN 1 END) AS cntPM,
COUNT(CASE WHEN inc_cat_id = 3 THEN 1 END) AS cntInstallation
FROM #incident
GROUP BY item_ID

Answer 3

UPDATE:
With the addition of logged_date column, you just need to add it in the ON clause. You need to use sp_executesql instead of EXEC now to prevent sql injection:
DECLARE #fromDate DATE = '2016-04-01',
#toDate DATE = '2016-04-30';
DECLARE #sql NVARCHAR(MAX) = N'';
SELECT #sql =
'SELECT
i.item_id' + CHAR(10) +
(SELECT
' , COUNT(CASE WHEN inc.inc_cat_id = ' + CONVERT(VARCHAR(10), inc_cat_id) +
' THEN 1 END) AS ' + QUOTENAME('count(' + inc_cat_n + ')') + CHAR(10)
FROM #Inc_cat
ORDER BY inc_cat_id
FOR XML PATH('')
) +
'FROM #item AS i
LEFT JOIN #incident AS inc
ON i.item_id = inc.item_id
AND inc.date_logged BETWEEN #fromDate AND #toDate
GROUP BY i.item_id;';
PRINT (#sql);
EXEC sp_executesql
#sql,
N'#fromDate DATE, #toDate DATE',
#fromDate,
#toDate
ONLINE DEMO
Giorgos answer is good if you only have that 3 Inc_cats. However, if you have unknown number of Inc_cats, you need to do it dynamically. Here is a method using a dynamic crosstab:
DECLARE #sql NVARCHAR(MAX) = N'';
SELECT #sql =
'SELECT
i.item_id' + CHAR(10) +
(SELECT
' , COUNT(CASE WHEN inc.inc_cat_id = ' + CONVERT(VARCHAR(10), inc_cat_id) +
' THEN 1 END) AS ' + QUOTENAME('count(' + inc_cat_n + ')') + CHAR(10)
FROM Inc_cat
ORDER BY inc_cat_id
FOR XML PATH('')
) +
'FROM Item AS i
LEFT JOIN Incident AS inc
ON i.item_id = inc.item_id
GROUP BY i.item_id;';
PRINT (#sql);
EXEC (#sql);
ONLINE DEMO
Basically, it's just a dynamic version of Giorgos' answer.