r - Difficultly filling NAs with imputed values by loop but not one-by-one -

so have matrix in r that's subset of dataframe holding little more id, grouping variable, , series of 1100 readings. i'm wanting run these readings through procedure requires spacing of values - each row (of approx 100) has 2 or 3 'missing values' day messing sensor various reasons. i'm wanting throw in loess imputation these values - 3-5% of each row generally, no more, i'm not terribly worried effect of few values other they're missing.

the code i'm using below. forgive inelegance in using loop lifting, , recognize yval not necessary i'm doing, makes easier track.

#take out matrix loess imputation df3mat <- as.matrix(cccsdf3[,c(3:1102)]) #generate list of time values loess generation timelist <- c(1:timevals) #use loess imputation fill values each row in df3mat matrix.  (i in nrow(df3mat)){   #take each row @ time yval vector   yval <- df3mat[i,]   #get vector of values in yval missing   missinglist <- which(is.na(yval))      #use span=0.13 loess curve, comes out 1 day of information in window fitting curve   yval.loess <- loess(y ~ x, span=0.13, data.frame(x=timelist, y=yval))   #get set of predictions missing values   yval.missing <-  predict(yval.loess, data.frame(x=missinglist))    #replace missing values in yval yval.missing objects imputed via loess   yval[is.na(yval)] <- yval.missing    #and set row in df3mat filled yval vector.   df3mat[i,] <- yval  } 

now, seem have happening this: when run code above, , view(df3mat), don't see changes @ in matrix final yval appears desired filled final row. however, if run 1 row (say, df3mat[6,]), works charm, , can see outcome in final df3mat. there problem in assignment? maybe hitting kind of memory problem in r?

you need fix for loop. error here:

for(i in nrow(df3mat)) {     ..... } 

nrow(df3mat) scalar value (single value) - not loop over. want like:

for(i in 1:nrow(df3mat)) {     ..... } 

or

for(i in seq_len(nrow(df3mat))) {     ..... }