Multiple IIFs and CHARINDEXs searches to extract data from same column in SQL? - sql

Question

This is MS SQL Server 2017.
This currently works. I just can't believe that this is the best way to perform these actions.
The Meeting table is populated from multiple services. It has a MeetingComment column that we used to encode some additional information from some of those sources but is unused by other sources. I want to extract some of the coded information, when used, into separate columns: Source and MeetingType.
MeetingComment
Source
MeetingType
{{{[ASTRA][CLASS][TMR:OFF]}}} ...
ASTRA
CLASS
{{{[ASTRA][CLASS][MERGED][TMR:OFF]}}} ...
ASTRA
CLASS
{{{[ASTRA][EVENT:Study Session][TMR:OFF]}}} ...
ASTRA
EVENT:Study Session
{{{[ASTRA][EVENT:Meeting][TMR:OFF]}}} ...
ASTRA
EVENT:Meeting
{{{[ASTRA][EVENT:Maintenance][TMR:ON]}}} ...
ASTRA
EVENT:Maintenance
UNK
UNK
Here is the SQL that I currently have that is working:
SELECT
Meeting.MeetingID,
Meeting.MeetingComment,
Meeting.Subject,
Rooms.RoomName,
IIF(
CHARINDEX(
'{{{[', Meeting.MeetingComment
) = 1,
SUBSTRING(
Meeting.MeetingComment,
5,
CHARINDEX(
']', Meeting.MeetingComment
)-5
),
'UNK'
) AS Source,
IIF(
CHARINDEX(
'{{{[', Meeting.MeetingComment
) = 1,
SUBSTRING(
Meeting.MeetingComment,
CHARINDEX(
'[',
Meeting.MeetingComment,
(
CHARINDEX(
'[', Meeting.MeetingComment,
1
)
) + 2
) + 1,
CHARINDEX(
']',
Meeting.MeetingComment,
(
CHARINDEX(
']', Meeting.MeetingComment,
1
)
)+ 2
)- CHARINDEX(
'[',
Meeting.MeetingComment,
(
CHARINDEX(
'[', Meeting.MeetingComment,
1
)
) + 2
) -1
),
'UNK'
) AS MeetingType,
Meeting.Recurrence,
Meeting.Location
But, as a programmer, it bugs me to have to use the same conditional test (the IIF statements) for both fields and to have to do the same CHARINDEX lookups multiple times. So before I move on, I just wanted to check to see if there is a better way to do this. Thanks in advance.

Answer 1

It appears from your example you are simply after the content of the first and second set of brackets.
Assuming the current version of SQL Server a more elegant solution would be a little parsing with json conbined with conditional aggregation:
select MeetingComment, Max([Source]) [Source], Max(MeetingType) MeetingType
from t
cross apply (
select
case when [key]='1' then j.[value] end [Source],
case when [key]='3' then j.[value] end MeetingType
from OpenJson(Concat('["', replace(Translate(MeetingComment,'[]',',,'), ',', '","'), '"]')) j
)s
group by MeetingComment;