r - For n rows, subtract one number; then next set of n rows switch to a new number to subtract from -

i have series of readings within 1 large dataframe shown below:

plate_id    day well_id name    x590_mean   x590_sd x750_mean   x750_sd 1   mcba15 001  0   spl1    water   0.196   0.003   0.145   0.004 2   mcba15 001  0   spl2    pyruvic acid methyl ester   0.202   0.001   0.143   0.000 3   mcba15 001  0   spl3    tween 40    0.214   0.036   0.158   0.026 4   mcba15 001  0   spl4    tween 80    0.196   0.000   0.144   0.002 5   mcba15 001  0   spl5    ?-cyclodextrin  0.217   0.012   0.161   0.012 ... 33  mcba15 001  1   spl1    water   0.209   0.008   0.111   0.003 34  mcba15 001  1   spl2    pyruvic acid methyl ester   0.371   0.007   0.148   0.003 35  mcba15 001  1   spl3    tween 40    0.481   0.127   0.285   0.088 36  mcba15 001  1   spl4    tween 80    0.242   0.011   0.108   0.002 37  mcba15 001  1   spl5    ?-cyclodextrin  0.277   0.002   0.138   0.001 

basically, need subtract mean values (590 , 750) in water rows each plate , day rows share same variables (ie. every 32 rows, switch next mean values subtract subsequent rows)

the desired output should be:

plate_id    day well_id name    x590_mean   x590_sd x750_mean   x750_sd 1   mcba15 001  0   spl1    water   0.000   0.003   0.000   0.004 2   mcba15 001  0   spl2    pyruvic acid methyl ester   0.006   0.001   0.000   0.000 3   mcba15 001  0   spl3    tween 40    0.018   0.036   0.013   0.026 ... 33  mcba15 001  1   spl1    water   0.000   0.008   0.000   0.003 34  mcba15 001  1   spl2    pyruvic acid methyl ester   0.162   0.007   0.037   0.003 

any values become negative should 0.

i have tried following approach run lot of difficulties.

sp2 <- split(dat, with(dat, interaction(plate_id, day))) sapply(sp2, dim) d <- function(biolog) {   x <- biolog$x590_mean[1]   biolog$x590_mean[biolog$x590_mean > x] <- biolog$x590_mean - x   y <- biolog$x750_mean[1]   biolog$x750_mean[biolog$x750_mean > y] <- biolog$x750_mean - y } lapply(sp2, d) 

you can try solution:

d <- function(biolog) {   biolog$x590_mean <- with(biolog, x590_mean - x590_mean[1])   biolog$x590_mean <- ifelse(biolog$x590_mean < 0, 0, biolog$x590_mean)    biolog$x750_mean <- with(biolog, x750_mean - x750_mean[1])   biolog$x750_mean <- ifelse(biolog$x750_mean < 0, 0, biolog$x750_mean)    biolog } 

(i haven't tested on actual data set though. please provide dput(dat) if there problems solution)