My data is like this
Name Values
A Val1
A Val1
A Val2
A Val2
A Val2
B Val1
B Val2
I want to ouput my data is this way
Name Val1Count Val2Count
A 2 3
B 1 1
I can get the Name and count(*) for Val1 with this query.
select [Name],count(*) FROM [table1]
where [values]='Val1'
group by [Name]
But I am not sure how to get the count(*) for val2 in the same row.
I tried doing this, but looks like this is not supported
select [name],#val1count= (above query for val1), #val2count = (above query for val2)
Please help. Thanks for looking.
This is called pivoting. Some databases provide a PIVOT function. However, you can also do this manually.
SELECT [Name],
SUM ( CASE WHEN [Values]='VAL1' THEN 1 ELSE 0 END ) AS Val1Count,
SUM ( CASE WHEN [Values]='VAL2' THEN 1 ELSE 0 END ) AS Val2Count
FROM [table1]
GROUP BY [Name]
Explanation:
The CASE WHEN ... END gives each row a "boolean" value for whether or not the row matches your condition.
The SUM ( ... ) counts the number of rows which returned "true" (or 1).
The GROUP BY [Name] consolidates the rows down to one row per distinct name.
If you add conditions to a WHERE clause, the CASE WHEN will only see the rows matching your WHERE conditions.
Related
I have a table with 4 columns, one of which is a non-null column, all others can be given a null value.
I just want a SELECT that returns results that have as many non-null columns as possible.
For example:
If I have 5 records in my Table, 1 of those records will have 3 columns with data, 2 of them will have 2 columns with data and 2 of them will have only 1 column with data. In my select, I want as a result only the first option: to bring me 3 columns with data. But it can be just 2 columns with data, it will be dynamic according to the table to be updated.
I'm using Oracle SQL.
One option is a conditional sort based on the count of non-null values in each row, and then a row-limiting clause (available since Oracle 12c). Assuming that the 3 nullable columns are col1, col2 and col3, that would be:
select *
from mytable
order by
case when col1 is null then 0 else 1 end
+ case when col2 is null then 0 else 1 end
+ case when col3 is null then 0 else 1 end desc
fetch first row with ties
If you are running an older version of Oracle, you can get the same result with window function rank():
select *
from (
select
t.*,
rank() over(order by
case when col1 is null then 0 else 1 end
+ case when col2 is null then 0 else 1 end
+ case when col3 is null then 0 else 1 end desc
) rn
from mytable t
) t
where rn = 1
I am quite new to SQL. Playing around with it and got stuck in the following scenario.
I have a table with the following data
My requirement is, in the FirstCol column, it has '11121' 3 times, and SecondCol has 1001 two times against FirstCol and it has 1002 against the same '11121' value which should not be. If it exists that way, I need to find all the data in such scenarios and need to display a column beside these two columns with a flag value indicating 1 if the value in the SecondCol is different than the other values. '0' should be displayed for the remaining.
I tried using group by, dense_rank() but couldn't get the desired results. Someone please help me out in getting the desired result for this.
I think you want:
select t.*,
(case when min(col1) over (partition by col2) =
max(col1) over (partition by col2)
then 0 else 1
end) as flag
from t;
This flags all rows where col2 has multiple values in col1.
You can use EXISTS and a correlated subquery, that checks, if there are row with the same firstcol but with a different secondcol. Put that in a CASE returning 1 if such a record exists, 0 otherwise.
SELECT t1.firstcol,
t1.secondcol,
CASE
WHEN EXISTS (SELECT *
FROM elbat t2
WHERE t2.firstcol = t1.firstcol
AND t2.secondcol <> t1.secondcol) THEN
1
ELSE
0
END flag
FROM elbat t1;
Here is the query
select col1
from table
col1 contains these category values:
A
B
C
NULL
How can I rename null category to D?
If you want to make the change permanent
UPDATE table
SET col1 = 'D'
WHERE col1 IS NULL
From then on you can simply query with ...
SELECT col1
FROM table
... to get the desired result.
If there is more than one row having a NULL in col1, you need to filter by a unique key, preferably by the primary key (which every table should have by the way). Let's say you have a table like
id (PK) col1
--- ----
1 'A'
2 'B'
3 'C'
4 NULL
5 NULL
then you can fix it with
UPDATE table SET col1 = 'D' WHERE id = 4;
UPDATE table SET col1 = 'E' WHERE id = 5;
unless you can calculate the new value from another column, e.g.:
UPDATE table
SET col1 = UPPER(LEFT(name, 1))
Try this : ISNULL( ) function is used to replace NULL value with another value
select isnull(col1,'D') as col1
from table
SQL Server uses ISNULL().
SELECT ISNULL(value_to_check, use_this_instead_if_valuetocheck_is_null)
For your code:
select ISNULL(col1, 'D') AS col_name
from table
However, this will happen across the board for this column. You can't use this to make a sequence, like D then E then F. Any NULL value you come across in this column will change to D.
I have a table with several columns and each column can have a value of 0 OR 1 ..I just want to eliminate the rows where all the columns have 0 .. I just want to consider the rows if at least one column have a value of 1
SELECT *
FROM A_TABLE_YOU_DID_NOT_NAME
WHERE col1 + col2 + ... + colN >= 1
Will not return all 0 and at least one of them will be 1.
IN SQL Server, you can make use of OR:
SELECT *
FROM table
WHERE [column1] = 1 OR [column2] = 1...
Assume we have the table as follows,
id Col-1 Col-2
A 1 some text
B 0 some other text
C 3
...
Take the table above as example, I want to build one SQL statement which would output the result: 2, 2.
The first value is the avg of all col-1 values except for 0, that is (1+3)/2 = 2. (If 0 is counted, then the result would be (1+0+3)/3 = 1, which is not what I want.)
The second value is the total number of all col-2 that is not empty. So the value is 2.
P.S, I know how to create them separately. What I prefer is to create only 1 statement to get both results.
For the first you can use NULLIF as null values are ignored in aggregations such as AVG.
For the second I assume you want to only count values not NULL or empty string.
SELECT AVG(NULLIF(Col1, 0)),
COUNT(CASE WHEN Col2 <> '' THEN 1 END)
FROM T
You want conditional aggregation:
select avg(case when col1 <> 0 then col1 end) as avg_not_zero,
count(col2) as num_not_empty
from table t;
As a note: 0 does not mean that the value is empty. Often NULL is used for this purpose in SQL, although strictly speaking, NULL means an unknown value.
Note: If "empty" could mean the empty string instead of NULL:
select avg(case when col1 <> 0 then col1 end) as avg_not_zero,
count(nullif(col2, '')) as num_not_empty
from table t;