AnalysisException error in hive using lag func - hive

I have a table which has some fields like name, id,errors, date_timestamp.
Name, id & date_timestamp are STRING, while errors is also STRING
Iam trying to get the previous value of errors field as below,
(lag(errors,1) over (partition by name, id order by date_timestamp) - errors) as err_calculate
I am getting the below error while executing the above code.
AnalysisException: Arithmetic operation requires numeric operands
I tried using cast() but no luck. Any help would be useful. Thanks.

Try the query below. If you're in the first row of the window, lag returns NULL and hive complains that it cannot do something like NULL - 1.
case when lag(errors,1) over (partition by name, id order by date_timestamp) is not null
then (lag(errors,1) over (partition by name, id order by date_timestamp) - errors)
else null
end as err_calculate

Try to find rows with not bigint values in errors you will see what is wrong with them
select * from ...
where errors is not null -- not null
and cast(errors as bigint) is null -- and not bigint
cast returns null when the value cannot be casted

Related

Can I know the issue of this SQL query

I have this SQL query:
but I'm getting an error:
If I remove the comma-separated value from the variable, it is working fine. As well as if I remove the NULL checking feature it is working fine. Can I know the issue of this
It's because a CASE WHEN can only return 1 value.
And a STRING_SPLIT returns a resultset.
I assume something like this is what you want.
SELECT *
FROM Facility f
WHERE (#Facility IS NULL OR f.facilityCode IN (SELECT value FROM string_split(#Facility,',')))
This will get all records if the variable is null.
The function dbo.split will split the string in more than one value. This will confuse your subquery and the error you are receiving will be thrown.
In case you need what goes before the comma consider using:
select top 1 value
from dbo.split(#Facility, ','))
You want to say if the variable is NULL so ignore the WHERE statement, if so your query would be:
SELECT *
FROM Facility f
WHERE #Facility IN (select value from dbo.split(#Facility, ',')) OR #Facility IS NULL

COUNTing IS NULL returns 0 or with an Error [BigQuery]

I want to find the number of data that is NULL but i do not know why this code returns an error: (Error Message said there was a problem with my WHERE)
SELECT first_review, COUNT (1) AS firstreviewisnull
FROM [table_name]
GROUP BY first_review
WHERE first_review IS NULL
I tried this code and the count returns to 0 - which I know is incorrect:
SELECT COUNT(first_review) AS firstreviewisnull
FROM [table_name]
WHERE first_review IS NULL
If possible, please help to:
explain what i did wrong
provide with the correct code
thank you!
SELECT COUNT(*) AS firstreviewisnull
FROM [table_name]
WHERE first_review IS NULL
Count() function ignores Null values (it does not count them). Use count(*) if you want to count the rows where girst_review is null.
Your SQL in the second image seems correct, assuming the [table_name] is actually replaced with the proper table name?
Also, check that the values for first_review are actually null, and not an empty String “”, which is not NULL.

Selecting from table to insert into another, getting a type error

I have the following query which inserts data into one table after selecting it from another.
The problem is that the data types do not match for one of the columns. I have simplified the query below.
INSERT INTO tbl.LogTable (
[SelPartNo], -- This does not match, see below
)
SELECT TOP 1
IF([SelPartNo] = 'False', NULL, [SelPartNo],
FROM tbl.MyTable
WHERE ID = '20358'
ORDER BY CreateDate DESC
The first SelPartNo is an int and the second is a VarChar. In most instances the SelPartNo for the second one (tbl.MyTable) is NULL or an integer, which I don't think will cause a problem. But in some cases the value is "False", which needs to return NULL.
I have tried an IF statement but I am doing something wrong because it's giving a syntax error and I am unsure if this is the correct approach.
Your code is syntactically incorect...
Try it with
NULLIF([SelPartNo],'False')
This function returns NULL if the two expressions are equal.
Details: https://msdn.microsoft.com/en-us/library/ms177562.aspx
I don't think IF is a function, at least not one which you can use in a SELECT statement. But CASE WHEN ... END is your friend:
INSERT INTO tbl.LogTable (
[SelPartNo]
)
SELECT TOP 1
CASE WHEN [SelPartNo] = 'False' THEN NULL ELSE [SelPartNo] END
FROM tbl.MyTable
WHERE ID = '20358'
ORDER BY CreateDate DESC

Check if field is numeric, then execute comparison on only those field in one statement?

This may be simple, but I am no SQL whiz so I am getting lost. I understand that sql takes your query and executes it in a certain order, which I believe is why this query does not work:
select * from purchaseorders
where IsNumeric(purchase_order_number) = 1
and cast(purchase_order_number as int) >= 7
MOST of the purchar_order_number fields are numeric, but we introduce alphanumeric ones recently. The data I am trying to get is to see if '7' is greater than the highest numeric purchase_order_number.
The Numeric() function filters out the alphanumeric fields fine, but doing the subsequent cast comparison throws this error:
Conversion failed when converting the nvarchar value '124-4356AB' to data type int.
I am not asking what the error means, that is obvious. I am asking if there is a way to accomplish what I want in a single query, preferably in the where clause due to ORM constraints.
does this work for you?
select * from purchaseorders
where (case when IsNumeric(purchase_order_number) = 1
then cast(purchase_order_number as int)
else 0 end) >= 7
You can do a select with a subselect
select * from (
select * from purchaseorders
where IsNumeric(purchase_order_number) = 1) as correct_orders
where cast(purchase_order_number as int) >= 7
try this:
select * from purchaseorders
where try_cast(purchase_order_number as int) >= 7
have to check which column has numeric values only.
Currently, in a table every field is setted with nvarchar(max) Like tableName (field1 nvarchar(max),field2 nvarchar(max),field3 nvarchar(3)) and tableName has 25lac Rows.
But on manually Check Field2 Contain the numeric Values Only... How to Check With t-sql that in the Complete Column (Field2) has numeric Value or not/null value with Longest Length in the Column!

Oracle SQL Case statement with NULL values

I'm trying to create an Oracle query that will use a parameter based on the CommandName of a button.
So here's my query
SELECT id
FROM table
WHERE dateTime IS (CASE WHEN :CommandName = 'FIRST' THEN NULL ELSE NOT NULL END)
So I'm passing the parameter but it's not working - I'm just getting a Missing NULL keyword error
I basically don't want to have to write two separate queries based on the parameter that is input
Not even sure if this is possible or if this is the right way of doing it?
Any ideas?
You may want to read up on the CASE statement a bit more.
I believe you want the following code:
SELECT
id
FROM
table
WHERE
(
(dateTime IS NULL AND :CommandName = 'FIRST')
OR
(dateTime IS NOT NULL AND NVL(:CommandName, '') <> 'FIRST')
)
Without using dynamic SQL to return the string 'NULL' or 'NOT NULL' you could use two NVL() functions:
select id
from table
where NVL(dateTime,'FIRST') = NVL(:param,dateTime);
However, if dateTime is indexed (and Oracle can index NULLs), ths NVL() function will disable the index.
I think that
SELECT id
FROM table
WHERE (dateTime IS NULL AND :CommandName='FIRST') OR (dateTime IS NOT NULL AND :CommandName <> 'FIRST')
should do what you need.