How to trim incomplete results from MySQL Select - mysql

Question

i have the next sub-query:
SELECT SUBSTRING(
(SELECT GROUP_CONCAT(DISTINCT PT.Factura SEPARATOR '|')
FROM darwin.vt_partidas PT
WHERE PT.Pedimento = P.ID)
,1,30) AS 'Resultado'
FROM darwin.vt_pedimentos P WHERE P.ID=130
I need to concat all results separated with | until i reach 130 characters, but my problem is that if at the end a result doesn't fit example:
i get the first 30 characters but the last result doesn't fit, i get:
result1|result2|result3|result
and i want this:
result1|result2|result3
(if the result doesn't fit, remove all characters from that result)
Thank you guys

Answer 1

Try this
Updated your GROUP_CONCAT and added another step to remove irrelevant data exceeding the 30 max length
SELECT
#str:= left(GROUP_CONCAT( DISTINCT PT.Factura SEPARATOR '|'), 30)
FROM
vt_pedimentos P
INNER JOIN
vt_partidas PT
ON PT.Pedimento = P.ID
WHERE
P.ID = 130;
-- to check whether the last or truncated text exists in the table otherwise remove
select
#str:= left(#str,
(
length(#str) - length(reverse(left(reverse(#str), locate('|', reverse(#str)) - 1)))
)
- 1)
FROM
vt_pedimentos P
where
NOT EXISTS
(
select
1
from
vt_partidas PT
where
PT.Factura = reverse(left(reverse(#str), locate('|', reverse(#str)) - 1))
)
and P.ID = 130;
further enhancement - have it to one sql statement

Answer 2

String manipulation is not a forte of SQL expressions.
But something like this should do it:
SELECT
IF(CHAR_LENGTH( GROUP_CONCAT(DISTINCT PT.Factura SEPARATOR '|') ) < 130
, GROUP_CONCAT(DISTINCT PT.Factura SEPARATOR '|')
, SUBSTRING_INDEX(
SUBSTR( GROUP_CONCAT(DISTINCT PT.Factura SEPARATOR '|') ,1,130)
, '|'
, CHAR_LENGTH( SUBSTR( GROUP_CONCAT(DISTINCT PT.Factura SEPARATOR '|') ,1,130) )
-CHAR_LENGTH(REPLACE(SUBSTR( GROUP_CONCAT(DISTINCT PT.Factura SEPARATOR '|') ,1,130),'|',''))
)
)
That's fairly complicated. It will be easier to decipher if we replace the GROUP_CONCAT expression with a placeholder. Let's have res represent GROUP_CONCAT(DISTINCT PT.Factura SEPARATOR '|') expression.
SELECT
IF(CHAR_LENGTH( res ) < 130
, res
, SUBSTRING_INDEX(
SUBSTR( res ,1,130)
, '|'
, CHAR_LENGTH( SUBSTR( res ,1,130) )
-CHAR_LENGTH(REPLACE(SUBSTR( res ,1,130),'|',''))
)
)
Still ugly, but better. Let's break that down.
If the number of characters in res is less than 130, we're done. Just return res.
Otherwise, we need to trim res to 130 characters, we can use SUBSTRING function to do that.
Now, we want to trim the last | and the following characters. To do that, we can get a count of the number | separator characters. Then we know which one the last one is.
(We can get a count of the separator characters by replacing all separator characters with an empty string, then getting the length of that string, and subtracting that from the length of the original string. The difference is the total length of the removed separator characters.
Then we can use that difference in a SUBSTRING_INDEX function to return all of the the characters before the last separator.
It's not a pretty solution. But it does implement an algorithm that satisfies the specification.