Conditional Where Statement - sql

I am trying to setup a conditional statement that will determine which WHERE clause to use.
I need to check 2 separate conditions, one on each side of the AND and if results are greater than 1 then use a particular statement if not then use nothing
Like this logic but in pdw sql
WHERE
if cte1.counte > 1 then 'cte1.built is not null' else ''
AND
if cte2.countd > 1 then 'cte2.demo is not null' else ''
possible combinations:
WHERE CTE1.BUILD IS NOT NULL
WHERE CTE1.BUILD IS NOT NULL AND CTE2.DEMO IS NOT NULL
WHERE CTE2.DEMO IS NOT NULL
BLANK
Is this possible to do?
Thanks in advance

Something like this:
WHERE (cte1.counte > 1 and cte1.built is not null or cte1.counte <= 1) and
(cte2.countd > 1 and cte2.demo is not null or cte2.countd <= 1)

Related

Two conditions failure in teradata case - NOT NULL AND NOT 0

Trying to return a column, giving 1, when column1 is NOT NULL and different than 0. So far managed to do this:
MAX(CASE WHEN column1 IS NOT NULL
THEN CASE WHEN column1 <> 0 THEN 1 ELSE 0 END
ELSE 0 END)
Getting this error:
SELECT Failed. 2620: The format or data contains a bad character.
It works quite ok with NOT NULL as a single condition, though.
I'm not sure why you need to nest anything, but it looks like you're missing the non-equality sign.
You might try,
CASE
WHEN column1 IS NOT NULL AND column1 <> 0
THEN 1
ELSE 0
END
Alternatively, this will produce the same result as an OR operator, where, but CASE executes the WHEN clauses in order.
CASE
WHEN column1 IS NOT NULL
THEN 1
WHEN column1 <> 0
THEN 1
ELSE 0
END
But in your description, it sounds like you wanted BOTH conditions to be true, so it doesn't make sense to nest or use multiple WHEN statements, because you can just connect them together with AND

Null validation in Hive , looking for best approach

I have condition in datastage which verifies 30 columns as not null and assigning value as '0' if anyone of the field is null like below,
isnull(columnA) or isNull(coloumnB) or isnull(columnC) or isNull(coloumnD) then 0 else 1.
we have option to use case statement in hive to set values, however we have to give use CASE statement per column like below,
select
case when (columnA is NULL) then 0
case when (columnB is NULL) then 0
case when (columnC is NULL) then 0
.
.
.
else 1
end as iValidColumn
From tablea
What i am looking?
Trying to look for option where we can validate all the 30 columns in condition for null validation.

CASE WHEN NULL makes wrong result in SQLite?

I have a table with a column of image type, the table has some rows but all the rows haven't had any image yet, they are all null. To test the CASE WHEN NULL, I've tried this and it gave a strange result:
SELECT CASE myImageColumn WHEN NULL THEN 0 ELSE 1 END FROM myTable
All the returned rows were in a column of 1's (I thought 0's). What is wrong here?
Your help would be highly appreciated!
Thank you!
You can't compare with NULL like that, you should try:
SELECT CASE WHEN myImageColumn IS NULL THEN 0 ELSE 1 END
FROM myTable
Use a different form of CASE instead:
SELECT CASE WHEN myImageColumn IS NULL THEN 0 ELSE 1 END FROM myTable
Two useful links:
http://www.sqlite.org/nulls.html
http://www.sqlite.org/lang_expr.html
There's a bypass:
CASE ifnull(myValue, 'someUniqueStringOrValue')
WHEN 'someUniqueStringOrValue' THEN 0 -- this means null
WHEN 'someNormalValue' THEN 1
END

Oracle SQL: Excluding a record from query results only when TWO conditions are met

So I have a pretty large Oracle SQL query. I want to add the following logic to my where clause, described in pseudocode below. Is it possible? I want to exclude a record if and only if BOTH columnA and columnB are null. If one or the other is null, that's okay.
IF (pfr.columnA && pfr.columnB != NULL)
exclude record
ELSE
do nothing
I tried the below except from my where clause, but obviously it does not accomplish what I need.
AND (pfr.columnA IS NOT NULL AND pfr.columnB IS NOT NULL)
where NOT (pfr.columnA is NULL and pfr.columnB is NULL);
WHERE
case when pfr.ColumnA is null then 0 else 1 end +
case when pfr.columnB is null then 0 else 1 end > 0
or
where (pfr.columnA is not null or pfr.columnb is not null)

How does 'in' clause works in oracle

select 'true' from dual where 1 not in (null,1);
when we execute this which will result nothing
what my question is:
is the above query is logically equivalent to
select 'true' from dual where 1 != null and 1 != 1;
which will result nothing just as above statement
Please clarify?
Correct (but note that IN is an operator, not a clause and it works like this in SQL in general, not only for Oracle).
where 1 not in (null,1)
is equivalent to:
where 1 != null and 1 != 1
which should really be written as:
WHERE 1 NOT IN (NULL, 1)
and
WHERE 1 <> NULL AND 1 <> 1
which is the same as:
WHERE (1 <> NULL) AND (1 <> 1)
which evaluates to:
WHERE UNKNOWN AND FALSE
and further as:
WHERE FALSE
So, it correctly returns no rows.
Notice that if you had WHERE 1 NOT IN (NULL, 2), it would evaluate to WHERE UNKNOWN (left as an exercise) and no rows would be returned either.
The issue of your script in comparing with NULL value. You should use
column is null and column = 1
Actually NULL is an undefined value. Any comparation with NULL gives neither True nor False but NULL. Even NULL = NULL
That's why your 1 not in (null,1) doesn't work.
Yes they are.
select something from table where column not in (1,2,3);
is equivalent to
select something from table where column != 1 and column != 2 and column != 3;
The IN statement is a collection of OR statements, while NOT IN is a collection of AND statements - but it is also not equal to.
So the NOT IN is equivalent to:
1 <> NULL
AND 1 <> 1
AND ...
While the IN would be equivalent to:
1 = NULL
OR 1 = 1
OR ...
Note that having NULL in the collection will not work, due to the quirky nature of NULL.
Yes. It is correct. Also NULL values should be compared with IS NULL