Hello, I have a dataset that looks like this: APP_ID QUARTER APP1 2019Q1 APP2 2019Q1 APP3 2019Q1 APP4 2019Q1 APP5 2019Q1 APP6 2019Q1 APP7 2019Q1 APP8 2019Q1 APP9 2019Q1 APP10 2019Q1 APP11 2019Q2 APP12 2019Q2 APP13 2019Q2 APP14 2019Q2 APP15 2019Q2 APP16 2019Q2 APP17 2019Q2 APP18 2019Q2 APP19 2019Q2 APP20 2019Q2 APP21 2019Q2 APP22 2019Q2 APP23 2019Q2 APP24 2019Q2 APP25 2019Q2 APP26 2019Q2 APP27 2019Q2 APP28 2019Q3 APP29 2019Q3 APP30 2019Q3 APP31 2019Q4 APP32 2019Q4 APP33 APP34 APP35 APP36 APP37 APP38 APP39 APP40 APP41 APP42 APP43 APP44 APP45 APP46 APP47 APP48 And the distribution of frequencies: QUARTER COUNT FREQ 2019Q1 10 20,8% 2019Q2 17 35,4% 2019Q3 3 6,3% 2019Q4 2 4,2% 16 33,3% I need to fill missing values with other quarters in such way that distribution of frequencies would be almost uniform (all values occur with almost the same frequency as much as possible). The number of application can change. As time goes in year, as the number can increase. Now there may appear applications with quarters 2019Q2 or 2019Q3 or missing values, but for example in december they can be application with 2019Q3 or 2019Q4 or missing values. Missing values can only be filled with quarters from the next year. The non-missing values cannot be changed. At this moment it could look like this: APP_ID QUARTER APP1 2019Q1 APP2 2019Q1 APP3 2019Q1 APP4 2019Q1 APP5 2019Q1 APP6 2019Q1 APP7 2019Q1 APP8 2019Q1 APP9 2019Q1 APP10 2019Q1 APP11 2019Q2 APP12 2019Q2 APP13 2019Q2 APP14 2019Q2 APP15 2019Q2 APP16 2019Q2 APP17 2019Q2 APP18 2019Q2 APP19 2019Q2 APP20 2019Q2 APP21 2019Q2 APP22 2019Q2 APP23 2019Q2 APP24 2019Q2 APP25 2019Q2 APP26 2019Q2 APP27 2019Q2 APP28 2019Q3 APP29 2019Q3 APP30 2019Q3 APP31 2019Q4 APP32 2019Q4 APP33 2019Q3 APP34 2019Q3 APP35 2019Q3 APP36 2019Q3 APP37 2019Q3 APP38 2019Q3 APP39 2019Q3 APP40 2019Q3 APP41 2019Q4 APP42 2019Q4 APP43 2019Q4 APP44 2019Q4 APP45 2019Q4 APP46 2019Q4 APP47 2019Q4 APP48 2019Q4 And then distribution of frequencies may be: QUARTER COUNT FREQ 2019Q1 10 20,8% 2019Q2 17 35,4% 2019Q3 11 22,9% 2019Q4 10 20,8% 0 0,0% I will be thankful for any suggestions. P.S. Sorry for my english.
... View more