I am using SQL Server 2014 and I have a column (called RankDetails) in my table. I need to extract 2 specific numbers which appear in those strings.
Rank Details column:
RankDetails
#1 of 223 hotels in Maldives
#3 of 223 hotels in Maldives
...
#10 of 223 hotels in Maldives
...
#126 of 223 hotels in Maldives
What I am looking for:
RankDetails Rank OutOf
#1 of 223 hotels in Maldives 1 223
#3 of 223 hotels in Maldives 3 223
... ... ...
#10 of 223 hotels in Maldives 10 223
... ... ...
#126 of 223 hotels in Maldives 126 223
I am able to extract the [Rank] column by using the following T-SQL code:
select
Rank = SUBSTRING([RankDetails], PATINDEX('%[0-9]%', [RankDetails]), PATINDEX('%[0-9][a-z !@#$%^&*(()_]%', [RankDetails]) - PATINDEX('%[0-9]%', [RankDetails]) 1),
*
from [MyTable]
However, I am having a hard time trying to figure out how to output the [OutOf] column.
I had a look here: SQL - How to return a set of numbers after a specific word for a possible solution but still cannot make it work.
CodePudding user response:
One of possible solutions is:
DECLARE @t TABLE (
Id INT NOT NULL PRIMARY KEY,
RankDetails VARCHAR(100) NOT NULL
);
INSERT INTO @t (Id, RankDetails) VALUES (1, '#126 of 223 hotels in Maldives');
INSERT INTO @t (Id, RankDetails) VALUES (2, '#10 of 223 hotels in Maldives');
WITH Words AS
(
SELECT Id,
Word = REPLACE(value, '#', ''),
WordNumber = ROW_NUMBER () OVER (PARTITION BY Id ORDER BY (SELECT 1))
FROM @T
CROSS APPLY STRING_SPLIT(RankDetails, ' ')
)
SELECT Id,
Rank = MAX(IIF(WordNumber = 1, Word, NULL)),
OutOf = MAX(IIF(WordNumber = 3, Word, NULL))
FROM Words
WHERE WordNumber IN (1, 3)
GROUP BY Id;
CodePudding user response:
It's a shame you haven't (yet) upgraded your version of SQL Server as newer versions make this much easier with functions like openJson.
From your sample data you can try the following:
select RankDetails,
Try_Convert(int, Replace(Left(rankdetails, NullIf(o, 0)-1), '#', '')) [Rank],
Try_Convert(int, Substring(rankdetails, o 3, l)) OutOf
from t
cross apply (values(CharIndex('of ', rankdetails) ))v(o)
cross apply (values(CharIndex('hotels', rankdetails) -o -4))h(l);
Example DB<>Fiddle
Just for fun a more modern version could work like this:
select RankDetails,
Max(case when k=0 then v end) [Rank],
Max(case when k=2 then v end) OutOf
from t
cross apply (
select Try_Convert(int,Replace(value,'#',''))v, [key] k
from OpenJson(Concat('["',replace(rankdetails, ' ', '","'),'"]'))
)x
group by RankDetails;