sorting and comparing columns in Big Query SQL - sql

Question

I have a requirement to to sort and compare columns values. In a table having 6 columns
Need to do sorting for A_Length, A_breadth, A_Width and similar sorting need to be done for B_length, B_breadth and B_width
After sorting comparison need to do be done between A_* and B_* column based on their sorting order like
After sorting:
comparison need to be done with out put true or False
(3<11=True and 22<22= false and 23<32 =true) over all result for this is false
(5<11=true and 11<22=true and 17<32=true ) over all result for this is true
(17<11=false and 23<22=false and 27<32=true) over all result for this is False
In Biq query i can do greatest and least but not sure how to take the 3rd value(that is neither greatest nor least)
Let me know if any one can suggest a logic Its a big table having multiple column and above 6 column will be part of it.
Some more info below :
Sorting is smallest to largest. suppose i have 6 columns with values: A_Len=3, A_Bred=15,A_Wid=10, B_Len=20, B_Bred=11,B_Wid=7 . So in this first sorting for A_* columns is needed (3,10,15) then sorting for B_* column needed(7,11,20). Then in same order comparison(less than) need to be done between A_* and B_* sorted values (3<7 = result True , 10<11 = result true ,15<20= result true), i need the output after comparison as true or false . Need an suggestion how this can be done in GCP BQ , all these 6 column are part of a table.
Regards,

Answer 1

It seems like you want do a comparison with something that is not quite the table you currently have. It seems that you want to sort each column individually, without affecting the others, and then compare.
You could try something like this (changing all the instances of TABLE for your table name):
SELECT
sorted_A_LENGTH.A_LENGTH,
sorted_A_BREADTH.A_BREADTH,
sorted_A_WIDTH.A_WIDTH,
sorted_B_LENGTH.B_LENGTH,
sorted_B_BREADTH.B_BREADTH,
sorted_B_WIDTH.B_WIDTH,
sorted_A_LENGTH.A_LENGTH < Sorted_B_LENGTH.B_LENGTH,
sorted_A_BREADTH.A_BREADTH < sorted_B_BREADTH.B_BREADTH,
sorted_A_WIDTH.A_WIDTH < Sorted_B_WIDTH.B_WIDTH
FROM
(SELECT A_LENGTH, row_number() OVER (ORDER A_LENGTH) AS row_num FROM TABLE ORDER BY ASC) as sorted_A_LENGTH
LEFT JOIN
(SELECT A_BREADTH, row_number() OVER (ORDER A_BREADTH) AS row_num FROM TABLE ORDER BY ASC) as sorted_A_BREADTH
ON sorted_A_LENGTH.row_num = sorted_A_BREADTH.row_num
LEFT JOIN
(SELECT A_WIDTH, row_number() OVER (ORDER A_WIDTH) AS row_num FROM TABLE ORDER BY ASC) as sorted_A_WIDTH
ON sorted_A_LENGTH.row_num = sorted_A_WIDTH.row_num
LEFT JOIN
(SELECT B_LENGTH, row_number() OVER (ORDER B_LENGTH) AS row_num FROM TABLE ORDER BY ASC) as sorted_B_LENGTH
ON sorted_A_LENGTH.row_num = sorted_B_LENGTH.row_num
LEFT JOIN
(SELECT B_BREADTH, row_number() OVER (ORDER B_BREADTH) AS row_num FROM TABLE ORDER BY ASC) as sorted_B_BREADTH
ON sorted_A_LENGTH.row_num = sorted_B_BREADTH.row_num
LEFT JOIN
(SELECT B_WIDTH, row_number() OVER (ORDER B_WIDTH) AS row_num FROM TABLE ORDER BY ASC) as sorted_B_WIDTH
ON sorted_A_LENGTH.row_num = sorted_B_WIDTH.row_num

Answer 2

Consider below
select *,
a_length < b_length as length_a_less_b,
a_breadth < b_breadth as breadth_a_less_b,
a_width < b_width as width_a_less_b
from (
select * from(select * from your_table limit 0) union all
select a_arr[offset(0)], a_arr[offset(1)], a_arr[offset(2)],
b_arr[offset(0)], b_arr[offset(1)], b_arr[offset(2)]
from your_table,
unnest([struct((
select array_agg(a order by a) from unnest([a_length, a_breadth, a_width]) a
) as a_arr)]),
unnest([struct((
select array_agg(b order by b) from unnest([b_length, b_breadth, b_width]) b
) as b_arr)])
)
if applied to sample data in your question
output is