Why can't I use my column alias in WHERE clause? - sql

Question

I want to compare a value of my current row with the value of the row before. I came up with this, but it won't work. It can't find PREV_NUMBER_OF_PEOPLE so my WHERE clause is invalid. I'm not allowed to use WITH. Does anyone have an idea?
SELECT
ID
,NUMBER_OF_PEOPLE
,LAG(NUMBER_OF_PEOPLE) OVER (ORDER BY DATE) AS PREV_NUMBER_OF_PEOPLE
,DATE
FROM (
SELECT * FROM DATAFRAME
WHERE DATE>=CURRENT_DATE-90
ORDER BY DATE DESC
) AS InnerQuery
WHERE NUMBER_OF_PEOPLE <> PREV_NUMBER_OF_PEOPLE

Answer 1

You have several issues with your query:
The filtering conditions should be in the outer query.
The new column definition should be in the inner query.
The order by should be in the outer query.
With these changes, it should work fine:
SELECT ID, NUMBER_OF_PEOPLE, PREV_NUMBER_OF_PEOPLE, DATE
FROM (SELECT D.*,
LAG(NUMBER_OF_PEOPLE) OVER (ORDER BY DATE) AS PREV_NUMBER_OF_PEOPLE
FROM DATAFRAME D
) AS InnerQuery
WHERE NUMBER_OF_PEOPLE <> PREV_NUMBER_OF_PEOPLE AND
DATE >= CURRENT_DATE - 90
ORDER BY DATE DESC;
You need the filtering after the LAG() so you can include the earliest day in the date range. If you filter in the inner query, the LAG() will return NULL in that case.
You need to define the alias in the subquery so you can refer to it in the WHERE. Aliases defined in a SELECT cannot be used in the corresponding WHERE. This is a SQL rule, not due to the database you are using.

Answer 2

You could use common table expression (CTE's) to split the query processing.
Something like this:
WITH cte1 AS
(
SELECT * -- field list is advised...
FROM DATAFRAME
WHERE DATE >= CURRENT_DATE-90
),
cte2 AS
(
SELECT ID
,NUMBER_OF_PEOPLE
,LAG(NUMBER_OF_PEOPLE) OVER (ORDER BY DATE) AS PREV_NUMBER_OF_PEOPLE
,DATE
FROM cte1
)
SELECT ID
,NUMBER_OF_PEOPLE
,PREV_NUMBER_OF_PEOPLE
,DATE
FROM cte2
WHERE NUMBER_OF_PEOPLE <> PREV_NUMBER_OF_PEOPLE
ORDER BY DATE DESC;

Answer 3

Logical query processing is the conceptual interpretation of the query that defines the correct result, and unlike the keyed-in order of the query clauses, it starts by evaluating the FROM clause. Understanding logical query processing is crucial for correct understanding of T-SQL.
The main statement used to retrieve data in T-SQL is the SELECT statement. Following are the main query clauses specified in the order that you are supposed to type them (known as “keyed-in order”):
SELECT
FROM
WHERE
GROUP BY
HAVING
ORDER BY
But as mentioned, the logical query processing order, which is the conceptual interpretation order, is different. It starts with the FROM clause. Here is the logical query processing order of the six main query clauses:
FROM
WHERE
GROUP BY
HAVING
SELECT
ORDER BY
You can use a CTE :
WITH CTE1 AS (
SELECT * FROM DATAFRAME
WHERE DATE>=CURRENT_DATE-90
),
CTE2 AS (
SELECT
ID
,NUMBER_OF_PEOPLE
,LAG(NUMBER_OF_PEOPLE) OVER (ORDER BY DATE) AS PREV_NUMBER_OF_PEOPLE
,DATE
FROM CT2
)
SELECT * FROM CT2
WHERE NUMBER_OF_PEOPLE <> PREV_NUMBER_OF_PEOPLE

Answer 4

Just move the lag() into the derived table.
SELECT *
FROM (
SELECT id,
number_of_people,
lag(number_of_people) over (order by date) as prev_number_of_people,
date
FROM dataframe
WHERE date >= current_date - 90
) AS InnerQuery
WHERE number_of_people <> prev_number_of_people
ORDER BY date DESC