How to lookup value higher in data.frame in R - r

Question

I have a time series data.frame where all values are below each other. But on every date there are more cases that come back regulary. Based on the time series I am adding a column with some calculations. These calculations are done case-specific. But for these calculations I need the value of the previous date of that case. I have now idea about which function to use. Can anybody point me to a function or an example somewhere on the net? Thanks!!
To be clear, this is what I mean. On date 1 the old value (before the score) for case 'a' is 1200. Based on the score of 1 the new value becomes 1250. On date 2 the I want this new value 1250 for placed in the column 'old value' (and than do some calculations to come to the new value, that new value has to be placed again in de column old value on date 4 or so et cetera). For case B the same. So the new value after the score on date 1 is 1190 and has to be placed in the correct row on date 3 (on date 2 there is now case B) et cetera for 1000's of cases and dates.
date name_case score old_value new_value
1 a 1 1200 1250
1 b 2 1275 1190
1 c 1 1300 1310
2 a 3 1250
2 c 1 1310
3 B 1 1190

Answer 1

Maybe this will do it. Assuming that we start with:
> dat
date name_case score old_value new_value
1 1 a 1 1200 1250
2 1 b 2 1275 1190
3 1 c 1 1300 1310
4 2 a 3 NA NA
5 2 c 1 NA NA
6 3 b 1 NA NA # note ... fixed cap issue
And then make a subset with values for new_value:
dat1 <- dat[ !is.na(dat$old_value), ]
And then replace the NA old_values with results from new_values in the subset by match-ing on name_case
dat[ is.na(dat$old_value) , "old_value" ] <-
dat1$new_value[ match(dat[ !is.na(dat$old_value) ,"name_case" ],
dat1$name_case)]
match generates a numeric vector that is used to index the new_values.
> dat
date name_case score old_value new_value
1 1 a 1 1200 1250
2 1 b 2 1275 1190
3 1 c 1 1300 1310
4 2 a 3 1250 NA
5 2 c 1 1190 NA
6 3 B 1 1310 NA