How to use multiple custom dimensions in Google Big Query - google-bigquery

Question

Is there a way to use multiple custom dimensions in GBQ without using the Max function? My problem of using Max function is that it only saves the max pax_num, but I would like to have the count of visitors for all of the combinations of ( Date,product.v2ProductCategory,eCommerceAction.action_type
,product.v2ProductName). Note the pax_num is number of pax on that ticket. I need every combination of the dest+pax_num, not the dest+max(pax_num)
SELECT
Date
,count(distinct( concat(FULLVISITORID,cast(visitID as string)))) as visitor
, product.v2ProductCategory as product_category
,max(if(customDimensions.index=2, customDimensions.value,null)) as dest
,max((if(customDimensions.index=21, customDimensions.value,null)) ) as pax_num
,eCommerceAction.action_type as Action_type
,product.v2ProductName as product_name
FROM `table` as t
CROSS JOIN UNNEST(hits) AS hit
CROSS JOIN UNNEST(hit.customDimensions) AS customDimensions
CROSS JOIN UNNEST(hit.product) AS product
GROUP BY
Date
,product.v2ProductCategory
,eCommerceAction.action_type
,product.v2ProductName

Answer 1

Not sure if this is what you are looking for, but if you include the field pax_num in the group by you might already find what you need, like so:
select
date,
count(distinct( concat(FULLVISITORID,cast(visitID as string)))) as sessions,
product.v2ProductCategory category,
max(if(customDimensions.index=2, customDimensions.value, null)) as dest,
if(customDimensions.index=21, customDimensions.value,null) as pax_num,
eCommerceAction.action_type as act_type,
product.v2ProductName as product_name
from `table` as t,
unnest(hits) as hit,
unnest(hit.customDimensions) customDimensions,
unnest(hit.product) as product
group by
date,
category,
act_type,
pax_num,
product_name
having pax_num is not null
You gave as an example the pax_num values "paxnum_5" and "paxnum_6". If you insert the value pax_num in the group by operation, the count aggregation should happen on the level of pax_num which would preserve the values (and not mix everything into the max value as before).
Also, notice that if you count the distinct combination of fullvisitorids and visitids you are actually computing the total amount of sessions and not visitors (their definition is not the same).

Answer 2

Add the fullvisitorID solve the problem
SELECT
Date
,concat(fullVisitorID,cast(visitID as string)) as visitorID
,count(distinct( concat(FULLVISITORID,cast(visitID as string)))) as visitor
, product.v2ProductCategory as product_category
,max(if(customDimensions.index=2, customDimensions.value,null)) as dest
,max((if(customDimensions.index=21, customDimensions.value,null)) ) as pax_num
,eCommerceAction.action_type as Action_type
,product.v2ProductName as product_name
FROM `table` as t
CROSS JOIN UNNEST(hits) AS hit
CROSS JOIN UNNEST(hit.customDimensions) AS customDimensions
CROSS JOIN UNNEST(hit.product) AS product
GROUP BY
Date
,product.v2ProductCategory
,eCommerceAction.action_type
,product.v2ProductName
,visitorID