I'm following along the steps in this article and am having trouble with the following query:
SELECT Quantity, COUNT(*) AS Quantity_Counts
FROM sqlbank3
WHERE UnitPrice >= 5
GROUP BY Quantity
HAVING Quantity_Counts < 450
ORDER BY Quantity_Counts DESC
LIMIT 10;
ERROR: column "quantity_counts" does not exist
LINE 5: HAVING Quantity_Counts < 450
I put everything in lower case to see if that did anything but to no avail. I have created Quantity as an integer. The query runs when I remove the HAVING clause so it's certainly finding the Quantity_Counts column, just not with the HAVING clause.
Any help is appreciated.
The HAVING clause is supposed to be evaluated before the SELECT clause. You cannot use the alias defined in SELECT in HAVING for that reason.
You must use
HAVING COUNT(*) < 450
instead.
(Some DBMS allow it to use alias names in the HAVINGclause, though.)
Related
I had a query made on aurora sql, it was working nice, but now I need to do the same in redshift, but when I do so, it throws an error asking me to group by by every column, but obviously I don't want that.
This is the query:
select
rut,
name,
id,
sum(cantidad_retornos) as cantidad_retornos,
sum(cantidad_aceptadas) as cantidad_aceptadas,
sum(cantidad_auto_accept) as cantidad_auto_accept,
sum(cantidad_rechazadas) as cantidad_rechazadas,
sum(cantidad_aceptadas) - sum(cantidad_auto_accept) as cantidad_aceptadas_manual,
coalesce((sum(cantidad_aceptadas) - sum(cantidad_auto_accept)) / nullif(sum(cantidad_aceptadas),0)) as per_aceptadas_manual,
coalesce(sum(cantidad_auto_accept) / nullif(sum(cantidad_aceptadas),0),0) as per_aceptadas_auto,
coalesce(sum(cantidad_rechazadas) / nullif(sum(cantidad_retornos),0),0) AS rechazo_per,
case
when coalesce(sum(cantidad_rechazadas) / nullif(sum(cantidad_retornos),0) ,0) < 0.1 or cantidad_retornos < 10 then 'Confiable'
when coalesce(sum(cantidad_rechazadas) / nullif(sum(cantidad_retornos),0),0) >= 0.1 and coalesce(sum(cantidad_rechazadas) / nullif(sum(cantidad_retornos),0),0) < 0.5 then 'Estándar'
when coalesce(sum(cantidad_rechazadas) / nullif(sum(cantidad_retornos),0),0) >= 0.5 then 'Poco confiable'
else 'Sin clasificar'
end as nivel_confianza
from table
where 1=1
group by id, name, rut
I tried to group by every column, but it doesn't throw the result that I need
The error that I get:
ERROR: column "reporte_sellers_date.cantidad_retornos" must appear in the GROUP BY clause or be used in an aggregate function
If I group by the third column, it throws the same error but with the column number 4
In the first option in the CASE statement you have or cantidad_retornos without any aggregating function such as SUM(). This is why Redshift is saying it needs to be in a group by. You also alias this name to the sum of the column of the same name. So the is a choice the database needs to make about which one to use - the source column or the aggregate. It looks like Aurora is choosing the aggregate but Redshift is choosing the source column.
Using the same name for an aggregate as a source column is not a good idea as you are relying on the query compiler to make a choice for you. This means the query can break when the compiler is updated or if you port the query to a different database.
To fix this you can either add the SUM() aggregation to the use of cantidad_retornos in the CASE statement or use the aggregate from above in the query but give it a unique name.
SELECT
Product_Line_ID=2 OR Product_Line_ID=3,
COUNT(Product_Finish), MIN(Standard_Price)
FROM Product_T
WHERE Product_Finish
GROUP BY Standard_Price
HAVING AVG(Standard_Price) <700
ORDER BY Product_FInish;
I keep getting this error: Your query does not include the specified expression 'Product_Line_ID=2 OR Product_Line_ID=3' as part of an aggregate function. Can anyone help me with this? Not sure how to select product line id that is 2 or 3.
What is confusing about the error message? Product_Line_ID=2 OR Product_Line_ID=3 is not valid SQL.
Your query basically makes no sense. You have boolean conditions in the SELECT, you have a WHERE clause with a column name and no conditions, you are ordering by the column you are counting.
I can guess that you intend something like this:
SELECT Product_Line_ID, COUNT(Product_Finish), MIN(Standard_Price)
FROM Product_T
WHERE Product_Line_ID IN (2, 3)
GROUP BY Product_Line_ID
HAVING AVG(Standard_Price) < 700
ORDER BY COUNT(Product_Finish);
I'm using PostgreSQL 10 and trying to run this query. I started with a CTE which I am referencing as 'query.'
SELECT
ROW_NUMBER()OVER() AS my_new_id,
query.geom AS geom,
query.pop AS pop,
query.name,
query.distance AS dist,
query.amenity_size,
((amenity_size)/(distance)^2) AS attract_score,
SUM((amenity_size)/(distance)^2) AS tot_attract_score,
((amenity_size)/(distance)^2) / SUM((amenity_size)/(distance)^2) as marketshare
INTO table_mktshare
FROM query
WHERE
distance > 0
GROUP BY
query.name,
query.amenity_size,
query.geom,
query.pop,
query.distance
The query runs but the problem lies in the 'markeshare' column. It returns the same answer with or without the SUM operator and returns one, which appears to make both the attract_score and the tot_attract_score the same. Why is the SUM operator read the same as the expression above it?
This is occurring specifically because each combination of columns in the group by clause uniquely identifies one row in the table. I don't know if this is intentional, but more normally, one would expect something like this:
SELECT ROW_NUMBER() OVER() AS my_new_id,
query.geom AS geom, query.pop AS pop, query.name,
SUM((amenity_size)/(distance)^2) AS tot_attract_score,
INTO table_mktshare
FROM query
WHERE distance > 0
GROUP BY query.name, query.geom, query.pop;
This is not your intention, but it does give a flavor of what's expected.
I'm trying to write a few Oracle SQL scripts for an assignment. I've managed to get all of it to work, except for one part. To summarize, I have to display data from 2 tables if the average of 1 column in table A is greater than the average of another column in table B. I realize you cannot include AVG functions in a WHERE clause or HAVING clause since it seems unable to properly access the data (from what I've read). When I exclude this clause, the script executes properly, so I'm confident there are no other errors.
I've tried writing it as follows but the error I get is ORA-00936: missing expression and it is just before the > sign. I thought this may be due to improper bracket placing but none of my attempts resolved this. Here is my attempt:
SELECT l.l_category, SUM(r.r_sold), AVG(l.l_cost)
FROM promos l
INNER JOIN sales r
ON r.promo_id = l.promo_id
GROUP BY l.l_category
HAVING (SELECT AVG(l.l_cost) OVER (PARTITION BY l.l_cost)) >
(SELECT AVG(r.r_sold) OVER (PARTITION BY r.r_sold));
I tried doing this without the OVER (PARTITION BY ...) as well as putting it into a WHERE clause but it didn't resolve the error. I'm pretty sure I need to put it into a SELECT statement somehow but I'm at a loss.
You do not need to use the OVER clause when applying the aggregate functions in the HAVING clause. Just use the aggregate functions on their own.
SELECT l.l_category, SUM(r.r_sold), AVG(l.l_cost)
FROM promos l
INNER JOIN sales r
ON r.promo_id = l.promo_id
GROUP BY l.l_category
HAVING HAVING AVG(l.l_cost) > AVG(r.r_sold)
I have the following query which returns the following error:
An aggregate may not appear in the WHERE clause unless it is in a subquery contained in a HAVING clause or a select list, and the column being aggregated is an outer reference.
SELECT Part from Parts Where count(Part) > 1
How could i rewrite it to return the part that appears more than once.
You need to use a GROUP BY and HAVING clause like this:
SELECT part
FROM Parts
GROUP BY part
HAVING COUNT(*) > 1
A perfect opportunity for the rarely used HAVING clause:
SELECT Part, Count(Part) as PartCount
FROM Parts
GROUP BY Part
HAVING Count(Parts) > 1
try this:
select part from parts group by part having count(part) > 1