SQL子字符串charindex

Question

I have a SQL field with a string like this

27/08/2018 to 31/08/2018 - GCH/10561//HO
04/09/2018 to 07/09/2018 - GCH/2836/202788279H/WAL

This string has a reference code I need to extract. The one bolded above.

The problem is the string before it has not always the same length. See second example above.

How can I extract that code each time?

Ive tried this, but i need to the left of the last "/".....

SELECT Right('27/08/2018 to 31/08/2018 - GCH/10561/201312361J/HO', CHARINDEX('/','27/08/2018 to 31/08/2018 - GCH/10561/201312361J/HO')-1)

I think I need to use a substring and charindex but cannot get it working.

Thanks in advance for your help.

Answer 1

This will locate the wanted string, but it requires quite a deal of reversing to get there. I have used a cross apply but you don't have to.

CREATE TABLE test(
   column_a VARCHAR(50) NOT NULL PRIMARY KEY
);
INSERT INTO test(column_a) VALUES ('27/08/2018 to 31/08/2018 - GCH/10561/201312361J/HO');
INSERT INTO test(column_a) VALUES ('04/09/2018 to 07/09/2018 - GCH/2836/202788279H/WAL');


select
      column_a, ca.x
from test
cross apply (
    select reverse(substring(reverse(column_a),CHARINDEX('/',reverse(column_a))+1,10))
    ) ca (x)

Answer 2

You can try below

DEMO

declare @text varchar(64)
set @text='27/08/2018 to 31/08/2018 - GCH/10561/201312361J/HO'
select reverse(substring(reverse(right(@text,CHARINDEX('-',@text)-3)),CHARINDEX('/',reverse(right(@text,CHARINDEX('-',@text)-3)))+1,10))

OUTPUT:

val
201312361J

Answer 3

This is assuming MS SQL Server:

declare @rev varchar(64)
set @rev=reverse('27/08/2018 to 31/08/2018 - GCH/10561/201312361J/HO')
SELECT 
reverse(SUBSTRING ( @rev,  charindex('/', @rev) + 1, charindex('/', substring(@rev, (charindex('/', @rev) + 1) + 1, Len(@rev)))));

Explanation:

• First we reverse the original string using reverse.
• Then we find the substring between the first occurrence of '/' and the next occurrence of '/'
• We reverse the string again to find our original string.

Answer 4

Assuming Postgres and that the - always precedes the second group of information:

with test(data) as (
values 
  ('27/08/2018 to 31/08/2018 - GCH/10561/201312361J/HO'), 
  ('04/09/2018 to 07/09/2018 - GCH/2836/202788279H/WAL')
)
select split_part(split_part(data, '-', 2), '/', 3)
from test;

returns:

split_part
----------
201312361J
202788279H