1

i hope, that here there is specialist of clementine or IBM SPSS MODELER. My question: How use Clem expression to transform date? For example i have this format of date

27.10.15.15:05
27.10.15.15:45
27.10.15.16:10

if the date value < 30 then, round to 1 min i.e. not 15:05, but 15:01 , if the date value > 30 min then, round to 31 min, i.e. not 15:45, but 15:31. What about 16:10, since it < 16:30 then, round to 16:01. generally speaking date must be in the follow format

27.10.15.15:01
27.10.15.15:31
27.10.15.16:01

One man told me, that it's should to use this expression datetime_in_seconds(?) But i really don't know how use it. Please help me to transform this date in needed for me format

eli-k
  • 10,898
  • 11
  • 40
  • 44
Hans Baxxter
  • 11
  • 3

1 Answers1

0

I resolved this problem. Maybe anyone needs in this desicion. Modeler has uncomfortable possibilities for data transformation

if (abs(time_mins_difference(datetime_time(time), datetime_time(datetime_hour(time), 1, 0))) < 15) then
datetime_timestamp(datetime_year(time), datetime_month(time), datetime_day(time), datetime_hour(time), 1, datetime_second(time))
else if (abs(time_mins_difference(datetime_time(time), datetime_time(datetime_hour(time), 1, 0))) > 15) and (abs(time_mins_difference(datetime_time(time), datetime_time(datetime_hour(time), 1, 0))) < 45) then 
datetime_timestamp(datetime_year(time), datetime_month(time), datetime_day(time), datetime_hour(time), 31, datetime_second(time))
else
datetime_timestamp(datetime_year(time), datetime_month(time), datetime_day(time), datetime_hour(time)+1, 1, datetime_second(time))
endif
endif
Hans Baxxter
  • 11
  • 3