Thank you again.
I did get your point, I am trying the same logic you suggested, I have created 2 rank nodes in one of them created rank_Column and in projection used it in calculated column as PREV_FAILURE = string("rank_Column" + 1)
on the second rank node created same rank_column and projection created calculated column as
FAILURE = "RANK_COLUMN" and both are NVARCHAR, and the rank properties: Ascending (Bottom N) Threshold = 5000, order by RPAR_DATE, partition by PART_CD.
and joined both projection from one projection I have renamed RPAR_DATE to PREV_FAIL_DATE and joined it to RPAR_DATE in the second projection and joined FAILURE to PREV_FAILURE but none of them is selected for output
I am getting DAYS_BETWEEN, and PREV_FAIL_DATE with PART_CD.
I am sure this is in the right direction.
At this point we are checking days in between RPAR_DATE and PREV_FAIL_DATE and if it's =< 90 days it will count it as a repeat, and this only counting consecutive days not all days in any given 90 days. Also we are calculating the repeat failure of this part which is universal, we need to calculate the repeat of a given part related to a specific serial number that's the only time it will count as repeat, when the same engine has same part replaced multiple times within 90 days.
So in this case do I add serial number as an attribute and join it between the 2 projection to achieve what I am looking for or you have a different approach?
Please advise and I thank you for your time.