提取字符串的两个部分

Question

One of the fields in my results, contains the following "standard" notes:

Audit Status changed from STOP to RELEASED.
Audit Status changed from STOP to RFS Pend.
Audit Status changed from RELEASED to RFS Pend.
Audit Status changed from STOP to ODY Pend.
Audit Status changed from ODY Pend to RELEASED.

There may be other notes but these are the ones that I have seen so far.

What I would like to do is to extract 2 parts of each note with an expression.

The first part is the words between from and to ( STOP, RELEASED, ODY Pend, etc. ).

The second part is the words after to ( RELEASED, RFS Pend, ODY, etc. )

I've tried split and it works but it makes the query run much longer (at least the way I tried to do for someone relatively new to sql). I've looked at using MID but I'm not sure how to write the expression since the starting point of each part is different.

Answer 1

This should do it with MS SQL Server.

CREATE TABLE MyTable (col1 int IDENTITY(1,1), col2 varchar(50));

INSERT MyTable (col2) VALUES ('Audit Status changed from STOP to RELEASED.');
INSERT MyTable (col2) VALUES ('Audit Status changed from STOP to RFS Pend.');
INSERT MyTable (col2) VALUES ('Audit Status changed from RELEASED to RFS Pend.');
INSERT MyTable (col2) VALUES ('Audit Status changed from STOP to ODY Pend.');
INSERT MyTable (col2) VALUES ('Audit Status changed from ODY Pend to RELEASED.');

And...

WITH MyData AS
(
  SELECT col1, 
    col2, 
    PATINDEX('% from %', col2) + 6 as Start1, 
    PATINDEX('% to %', col2) as Stop1,
    PATINDEX('%.', col2) as Stop2
  FROM MyTable
)
SELECT SUBSTRING(col2, Start1, Stop1-Start1) as Word1,
       SUBSTRING(col2, Stop1+4, Stop2-Stop1-4) as Word2
FROM MyData

I can't say that the performance will be any better than what you're doing now, but you can try it out and let us know.

NOTE: This will break down if the patterns (' from ', ' to ' or '.') occur more than once in your strings. And you'll want to include some error handling or maybe a where clause to filter out values of zero (when the pattern is not found) that will result in negative lengths.