Convert WHERE logic to SQL statement - sql

Can someone help me transform the following code to raw SQL statement?
(NOT to Dynamic SQL)
Dim blnAllow as Boolean = True
Dim intType as Int32 = 35
.Append("SELECT * FROM TABLE1 WHERE NAME='AAA' ")
Select Case intType
Case 35
.Append("AND (Type IN (2,4) OR type=8) ")
.Append("AND [use]=1 ")
Case 34
If blnAllow = True Then
.Append("AND (Type IN (2,4) OR (type=8 and Col1 > 0 )) ")
Else
.Append("AND (Type IN (2,4)) ")
End If
.Append(" AND [use]=1 ")
Case Else
.Append("AND Type=1")
End Select

Well since intType is defined as 35, only the Case 35 section applies...
select * from TABLE1 where [NAME]='AAA'
and [Type] in (2,4,8)
and [use] = 1
If you want to encapsulate those other cases, you'll have to explain where intType fits in.. or do you just want 3 separate queries?

How about something like this
SELECT *
FROM TABLE1 WHERE NAME='AAA'
AND (
(
intType = 35
AND (Type IN (2,4) OR type=8)
AND [use]=1
)
OR
(
intType = 34
AND (
(
blnAllow = 'true'
AND (Type IN (2,4) OR (type=8 and Col1 > 0 ))
)
OR
(
blnAllow = 'false'
AND (Type IN (2,4))
)
)
AND [use]=1
)
OR
(
intType NOT IN (35, 34)
AND Type=1
)
)

The general converting pattern is
If condition Then ands1
Else ands2 End If
becomes
( (condition AND ands1) OR ((NOT condition) AND ands2) )

Please try this most optimized query.
select * from table where name = 'AAA' AND
(
(
((Type IN (2,4) OR type=8) OR // Case 35
(
(Type IN (2,4) OR (type=8 and Col1 > 0 )) // Case 34 and blnAllow checking
)
)
AND [use]=1 // Case 35 && 34
) OR
(Type=1) // Else
)
If the string "Type" and "type" indicates the same field, You just modifify the
//case 35 section to (Type IN (2,4) OR type=8) => (Type IN (2,4,8))

In MS SQL, this would look like this:
DECLARE #blnAllow BIT
SET #blnAllow = 1
DECLARE #intType INT
SET #intType = 35
SELECT *
FROM TABLE1
WHERE
NAME = 'AAA' AND
(
(#intType = 35 AND (Type IN (2,4) OR type = 8) AND [use] = 1) OR
(#intType = 34 AND [use] = 1 AND
(
(#blnAllow = 1 AND (Type IN (2,4) OR (type = 8 and Col1 > 0 ))) OR
(#blnAllow = 0 AND (Type IN (2,4)))
)) OR
(#intType not in (34, 35) AND Type = 1)
)
Don't expect query optimizer to optimize it :).

…
WHERE NAME = 'AAA'
AND (#intType NOT IN (34, 35) AND Type = 1
OR #intType IN (34, 35) AND [use] = 1 AND (
Type IN (2, 4)
OR #intType = 35 AND Type = 8
OR #intType = 34 AND (#blnAllow = 0 OR Type = 8 AND Col1 > 0)
)
)
It is assumed that #intType is an int parameter and #blnAllow is a bit parameter.

Related

Filter only get one condition

I have a validation in WHERE clause like:
...
AND (#BDOnly = 0
OR [DT].[Abbreviation] = #IsBDChecked)
AND (#CDOnly = 0
OR [DT].[Abbreviation] = #IsCDChecked)
AND (#PPOOnly = 0
OR [DT].[Abbreviation] = #IsPPOChecked)
AND (#FBOMOnly = 0
OR [DT].[Abbreviation] = #IsFBOMChecked)
AND (#APOnly = 0
OR [DT].[Abbreviation] = #IsAPChecked)
AND (#COOnly = 0
OR [DT].[Abbreviation] = #IsCOChecked)
So each AND clause check if boolean is bit value is 0 or 1, if it's 0 just do any but if it's 1 it do a filter. My problem is
if I'm sending two with value 1, I mean
#BDOnly = 1 and #CDOnly = 1 it only filter one of them instead get two filters I mean:
[DT].[Abbreviation] = #IsBDChecked
and
[DT].[Abbreviation] = #IsCDChecked
What am I doing wrong? Regards
I'm going to take a guess here because there's not enough data. If you want to return a row where Abbreviation matched #IsBDChecked only when #BDOnly is set, and also another row where Abbreviation matches #IsCDChecked when #CDOnly is set, try this:
(#BDOnly = 1
AND [DT].[Abbreviation] = #IsBDChecked)
OR (#CDOnly = 1
AND [DT].[Abbreviation] = #IsCDChecked)
OR (#PPOOnly = 1
AND [DT].[Abbreviation] = #IsPPOChecked)
OR (#FBOMOnly = 1
AND [DT].[Abbreviation] = #IsFBOMChecked)
OR (#APOnly = 1
AND [DT].[Abbreviation] = #IsAPChecked)
OR (#COOnly = 1
AND [DT].[Abbreviation] = #IsCOChecked)
Small representative case:
DECLARE #ADOnly BIT = 0
, #BDOnly BIT = 1
, #CDOnly BIT = 1
, #IsADChecked NVARCHAR(100) = 'A'
, #IsCDChecked NVARCHAR(100) = 'C'
, #IsBDChecked NVARCHAR(100) = 'B'
;WITH myTable AS
(
SELECT * FROM (VALUES ('A'), ('B'), ('C')) X(Abbreviation)
)
SELECT *
FROM myTable
WHERE (#ADOnly = 1 AND Abbreviation = #IsADChecked)
OR (#BDOnly = 1 AND Abbreviation = #IsBDChecked)
OR (#CDOnly = 1 AND Abbreviation = #IsCDChecked)
Yields:
Abbreviation
------------
B
C

Using IF Statement inside a Where Clause in SQL Server

I have two params that are passed in a function which passed down to the SQL string. Variables can be null or has a value (int). If x is not null, use "this" column else use "that" column. I'm using SQL Server.
// Inside a function with two variables passed, x and y
$sql = "
SELECT
[...]
FROM
[...]
WHERE
[...]
AND [...]
-- This is the tricky part
AND
--- if $x is not null, use foo column else use bar column
IF (x, R.column = 2, R.another_column = 3)
[...]
";
Is this possible to select a column based on the value of the variable passed in?
AND
(
($x is not null and R.column = 2) OR
($x is null and R.another_column = 3)
)
Unless I'm misunderstanding, you just need logic, or a case statement would work.
$sql = "
SELECT
[...]
FROM
[...]
WHERE
[...]
AND
[...]
-- This is the tricky part
AND
--- if $x is not null, use foo column else use bar column
-- IF (x, R.column = 2, R.another_column = 3)
(
(X IS NULL AND R.column = 2)
OR
(X IS NOT NULL AND R.another_column = 3)
)
";
AND
--- if $x is not null, use foo column else use bar column
case
when x is NULL then R.another_column = 3
else R.column = 2
end
--in SQL Server...
DECLARE #x INT = 3;
SELECT *
FROM sys.schemas
WHERE (#x IS NULL AND schema_id = 2)
OR (#x IS NOT NULL and schema_id = 3)
--optional depending on how much you execute this query, it may help
OPTION(RECOMPILE)

Datatables combine data from different columns into single row

Refer to the below image, how to merge all chapter into one single rows of different page number?
e.g.
Page 1 has 3 rows as there are 3 data at columns of chp 8,9,10
Page 2 has 5 rows as there are 5 data at columns of chp 8,9,1,11,12
So
how to have a single row of each page number and the all data placed at their own chp?
Basic SQL as below:
SELECT `mc_ee_page_id` as Page,
IF ( mc_ee_chp_id = 1, ( CASE
WHEN (is_marked = 0 and mk_no_mark = 1 ) THEN "Y"
WHEN (is_marked = 1 and mk_no_mark = -1) THEN "N"
WHEN (is_marked = 1 and mk_no_mark = 1 ) THEN "O" ELSE "X"
END ) , "" ) AS chp_1 ,
IF ( mc_ee_chp_id = 2, ( CASE
WHEN (is_marked = 0 and mk_no_mark = 1 ) THEN "Y"
WHEN (is_marked = 1 and mk_no_mark = -1) THEN "N"
WHEN (is_marked = 1 and mk_no_mark = 1 ) THEN "O" ELSE "X"
END ) , "" ) AS chp_2 ,
IF ( mc_ee_chp_id = 3, ( CASE
WHEN (is_marked = 0 and mk_no_mark = 1 ) THEN "Y"
WHEN (is_marked = 1 and mk_no_mark = -1) THEN "N"
WHEN (is_marked = 1 and mk_no_mark = 1 ) THEN "O" ELSE "X"
END ) , "" ) AS chp_3 ,
:
:
IF ( mc_ee_chp_id = 12, ( CASE
WHEN (is_marked = 0 and mk_no_mark = 1 ) THEN "Y"
WHEN (is_marked = 1 and mk_no_mark = -1) THEN "N"
WHEN (is_marked = 1 and mk_no_mark = 1 ) THEN "O" ELSE "X"
END ) , "" ) AS chp_12
FROM `maxcare_mc_ee_hw`
WHERE `stud_id` = '3312' AND `mc_ee_level_id` = '1'
GROUP BY mc_ee_chp_id, mc_ee_page_id
ORDER BY mc_ee_page_id asc
LIMIT 500
One way would be to GROUP the results of your SQL by making it the subquery of an outer query...
SELECT Page, MAX(chp_1) AS chp_1, MAX(chp_2) AS chp_2, MAX(chp_3) AS chp_3, MAX(chp_4) AS chp_4, MAX(chp_5) AS chp_5, MAX(chp_6) AS chp_6, MAX(chp_7) AS chp_7, MAX(chp_8) AS chp_8, MAX(chp_9) AS chp_9, MAX(chp_10) AS chp_10, MAX(chp_11) AS chp_11, MAX(chp_12) AS chp_12
FROM
(
-- *** paste your SQL in here ***
) AS indivdual_rows
GROUP BY Page
It assumes a page cannot have more than one non-blank value for the same chapter. If that is not the case, then it will only show the MAX alphabetical value.
Click here to go to SQL Fiddle to see it in action.

RunningSum inside another

I have a report with the following data
IdSource 20170629 20170628 20170626 20170625 20170624 20170623
Id1. OK KO N/A KO OK KO
I want to count the number of days my data (status of a workflow) is KO. N/A means the worflow is not expected, so it should not be counted.
Expected results :
IdSource 20170629 20170628 20170626 20170625 20170624 20170623
Id1. OK KO(2) N/A KO(1) OK KO(1)
The count must be reset at each OK.
In SQL, I would do
select t.*,
sum(case when status = 'KO' then 1 else 0 end) over (partition by id, cume_ko order by date) as nbDayKO
from (select t.*,
sum(case when status = 'OK' then 1 else 0 end) over (partition by id order by date) as cume_ko
from t
) t
I tried to use the RunningSum function without success:
I need to inverse the sort on the date
I cannot use the result of a runningSum inside another.
Thanks for your help
It's ugly and not very efficient, but I did not find any other solutions. Here is how I did it :
If ([status] = 0) Then 0
Else (
(If (RelativeValue([status];([dayOfWeek]);1) = 1) Then
If (RelativeValue([status];([dayOfWeek]);2) = 1) Then
If (RelativeValue([status];([dayOfWeek]);3) = 1) Then
If (RelativeValue([status];([dayOfWeek]);4) = 1) Then
If (RelativeValue([status];([dayOfWeek]);5) = 1) Then
If (RelativeValue([status];([dayOfWeek]);6) = 1) Then
If (RelativeValue([status];([dayOfWeek]);7) = 1) Then
If (RelativeValue([status];([dayOfWeek]);8) = 1) Then
If (RelativeValue([status];([dayOfWeek]);9) = 1) Then
If (RelativeValue([status];([dayOfWeek]);10) = 1) Then
If (RelativeValue([status];([dayOfWeek]);11) = 1) Then
If (RelativeValue([status];([dayOfWeek]);12) = 1) Then
If (RelativeValue([status];([dayOfWeek]);13) = 1) Then
If (RelativeValue([status];([dayOfWeek]);14) = 1) Then
If (RelativeValue([status];([dayOfWeek]);15) = 1) Then 15
Else 14
Else 13
Else 12
Else 11
Else 10
Else 9
Else 8
Else 7
Else 6
Else 5
Else 4
Else 3
Else 2
Else 1
Else 0)+[status])

Rewrite subquery with calculation in SQL to CASE statement

I have, as a part of a bigger query, some subqueries that I would like to convert to CASE statements instead.
The subquery looks like this (and works):
(SELECT (((SUM(DAm)-(SUM(StcCst)*-1))*100)/NULLIF(SUM(DAm),0)) AS 'DG' FROM [F0001].[dbo].[ProdTr] WHERE AcYrPr = '201601' AND ProdTr.TrTp = 1 AND [F0001].[dbo].[ProdTr].CustNo = '12773') AS dg_period_1
However, I don't seem to find any logical way to put this into a CASE-statement.
Any help would be appreciated!
(
SELECT CASE
WHEN SUM(t1.DAm) <> 0
THEN (SUM(t1.DAm) + SUM(t1.StcCst)) * 100 / SUM(t1.DAm)
ELSE 0 /* or whatever you want to have in this case */
END AS 'DG'
FROM [F0001].[dbo].[ProdTr] t1
WHERE t1.AcYrPr = '201601' AND
t1.TrTp = 1 AND
t1.CustNo = '12773'
) AS dg_period_1
I also removed some unneeded parentheses and simplified an operation (x - (y * -1) = x + y)
You could use the following statement with CASE provided you want to return a null when SUM(DAm) is null or 0.
(SELECT CASE
WHEN SUM(DAm) IS NOT NULL and SUM(DAm) <> 0 THEN (((SUM(DAm) - (SUM(StcCst) * -1)) * 100) /SUM(DAm))
ELSE NULL
END AS 'DG'
FROM [F0001].[dbo].[ProdTr]
WHERE AcYrPr = '201601'
AND ProdTr.TrTp = 1
AND [F0001].[dbo].[ProdTr].CustNo = '12773') AS dg_period_1