find max values

colinr

Hello,

 

I have data that looks like this:

 

a,1

a,2

a,3

b,1

b,2

c,1

c,2

c,3

c,4

d,1

e,1

f,1

f,2

 

I want to add a beast mode field so that the data in that field displays like this.  In this way the beast mode field would return the highest value from counting the first two fields.  

 

a,1,-

a,2,-

a,3,3

b,1,-

b,2,2

c,1,-

c,2,-

c,3,-

c,4,4

d,1,1

e,1,1

f,1,-

f,2,2

 

I hope that makes sense!  I'm a beginner in these statements having made a few case and date statements. Appreciate any available input or direction ofn where to start!

ST_Superman

This is not something that you can accomplish with a beast mode.  

 

I did this with a windowed select statement using a Redshift DataFlow:

 

select
"type"
,"number"
,MAX("number") over (partition by "type") as "max"  " -- this will add a column with the max "number" value for each "type"
from dojo_help
order BY "type" ASC

 

I saved this as max_values and then did this transform to match your desired output:

 

SELECT
"type"
,"number"
,case
when "number" = "max" then cast("number" as CHAR) -- in order to have '-' in the rows that do not have a max value you need to cast "number" as a CHAR
else '-' end as "highest value"
FROM max_values
order by "type" ASC, "number" ASC

 

 

ST_Superman

This is not something that you can accomplish with a beast mode.  

 

I did this with a windowed select statement using a Redshift DataFlow:

 

select
"type"
,"number"
,MAX("number") over (partition by "type") as "max"  " -- this will add a column with the max "number" value for each "type"
from dojo_help
order BY "type" ASC

 

I saved this as max_values and then did this transform to match your desired output:

 

SELECT
"type"
,"number"
,case
when "number" = "max" then cast("number" as CHAR) -- in order to have '-' in the rows that do not have a max value you need to cast "number" as a CHAR
else '-' end as "highest value"
FROM max_values
order by "type" ASC, "number" ASC

 

 

colinr

Many thanks!