Get COUNT() of rows where first 3 digits of column are alike - sql-server

Question

I have a result set of codes that are usually three digits followed by up to 2 digits like 012.34 or 123.45. The first three digits define a general category of group, and the digits following the decimal place define more specific qualities. There could be 77 012.xx numbers, and there are hundreds of unique 3 digit group definitions, followed by a varying number of digits per entry.
Does anyone know how to write a quick query to achieve this?

Answer 1

Assuming it is in a varchar column since you're storing 012.34...
SELECT LEFT(someColumn,3), COUNT(*)
FROM someTable
GROUP BY LEFT(someColumn,3)
HAVING COUNT(*) > 5 -- per your comments
ORDER BY LEFT(someColumn,3)
If it's not, then you'd do this:
SELECT LEFT(CONVERT(VARCHAR(10),someColumn),3), COUNT(*)
FROM someTable
GROUP BY LEFT(CONVERT(VARCHAR(10),someColumn),3)
ORDER BY LEFT(CONVERT(VARCHAR(10),someColumn),3)

Answer 2

#rypress these look strikingly similar to ICD-9 Diagnosis codes for Other respiratory tuberculosis ClickMe. Is this correct? In that case you will get Category and subcategory in your result and may change your counts(012->012.0->012.00,012.02..)
Sample Data:
IF OBJECT_ID(N'TempICD') > 0
BEGIN
DROP TABLE TempICD
END
CREATE TABLE TempICD (ICD VARCHAR(10))
INSERT INTO TempICD
VALUES ('012'),('012.0'),('012.00'),('012.01'),('012.02'),
('012.03'),('012.04'),('012.05'),('012.05'),
('013'),('013.0'),('013.00'),('013.01'),('013.02'),
('013.03'),(NULL)
Query to get Category with 6 or more line items (Including Category and Sub Category):
SELECT LEFT(ICD, 3) AS ICDs,
COUNT(1) AS ICDCount
FROM TempICD
GROUP BY LEFT(ICD, 3)
HAVING COUNT(*) > 5
ORDER BY LEFT(ICD, 3)
Query to get Category with 6 or more line items (Excluding Category and Sub Category):
SELECT SUBSTRING(ICD, 1, CHARINDEX('.', ICD + '.') - 1) AS ICDs,
SUM(CASE
WHEN LEN(SUBSTRING(ICD, CHARINDEX('.', ICD) + 1, LEN(ICD))) = 2 THEN 1
ELSE 0
END) AS ICDCount
FROM TempICD
WHERE ICD IS NOT NULL
GROUP BY SUBSTRING(ICD, 1, CHARINDEX('.', ICD + '.') - 1)
HAVING SUM(CASE
WHEN LEN(SUBSTRING(ICD, CHARINDEX('.', ICD) + 1, LEN(ICD))) = 2 THEN 1
ELSE 0
END) > 5
Cleanup Script:
IF OBJECT_ID(N'TempICD') > 0
BEGIN
DROP TABLE TempICD
END

Answer 3

This may also help you.
I assume that the column is Decimal Data type.
SELECT CAST([COLUMN] AS INT) [GROUP],
COUNT(*) [COUNT]
FROM [TABLE] T
GROUP BY CAST([COLUMN] AS INT)

Answer 4

it depends if your resultset is numeric or characters. if not numeric you can use string operations.
select left(resultSet,charindex('.',resultset)) , count(*)
from x
group by by left(resultSet,charindex('.',resultset))
order by left(resultSet,charindex('.',resultset))
with charindex you will get a correct 'cut' when the first digits are not 3 as 'usually'.
if your resultset is numeric/float you can use the floor function
select floor(resultset),count(*)
from x
group by floor(resultset)
order by floor(resultset)