How can I replace multiple values per cell with a mean per cell for multiple columns and rows in R

Question

I have a data set that I need to tidy, but I cannot figure out how to do it in an efficient way. The data set has multiple values per cell in multiple columns, and I want to replace the multiple values per cell with one value, preferably the mean of the multiple values.

I am thinking that I need to first separate the columns with multiple values per cell into multiple columns (so that there is only one value per cell) and then calculate the mean and get that mean into a new column replacing the original column (and the extra columns I created while separating). How would I do this in an efficient way? Or what is a better solution?

My data has 663 rows and 55 columns and many of the columns have multiple values per cell. Here is an example of parts my data:

                   X h_max h_min       seed_wght ell_light_uk ell_moist_uk ell_pH_uk  ell_N  ell_S
1    Achillea millefolium    45     8        0.16;0.2        7;8;7        5;5;4   6x;6;NA  4;5;5  1;1;1
2       Achillea ptarmica    60    20        0.31;0.2        7;8;7        8;7;8     4;5;4  3;2;4  0;0;0
3   Aegopodium podagraria   100    40        1.99;2.2       5;6x;6        6;5;6     6;7;7  8;8;7  0;0;1
4     Agrimonia eupatoria    60    30         2.96;13        7;7;7        4;4;4     8;7;8  4;4;4  0;0;1
5         Agrostis canina    70    10       0.05;0.06        7;7;9        7;9;9     3;3;3  2;3;2  0;0;0
6     Agrostis capillaris    70    10 0.07;0.06;0.075        7;6;7      4x;NA;5     4;4;4  4;4;4  0;0;0
7       Agrostis gigantea    70    10       0.08;0.09        7;7;7        6;6;8     6;7;7  6;6;7  0;1;0
8    Agrostis stolonifera   100    15       0.07;0.02        7;7;8       6;7~;7   7;6x;NA  5;6;5  1;0;1
9       Ajuga chamaepitys    20     5     1.6602;1.16        7;7;8        4;3;4     8;9;8  2;2;4  0;0;0
10       Ajuga genevensis  <NA>  <NA>            <NA>         <NA>         <NA>      <NA>   <NA>   <NA>
11      Ajuga pyramidalis    30    10       1.373;1.6        7;7;7        5;5;5     1;1;5  1;2;2  0;0;0
12          Ajuga reptans    30    10             1.4        5;5;6        6;7;6     6;5;6  5;6;6  0;0;0
13      Alchemilla glabra    60  <NA>            0.51       7;6;NA       5;NA;6    6;6;NA 4;NA;6 0;0;NA
14 Alchemilla glaucescens    20  <NA>      0.487;0.46        7;7;7        5;5;5     4;4;7  3;5;4  0;0;0
15   Alchemilla monticola    40  <NA>           0.689        6;7;6        5;4;5     6;6;6  4;4;4  0;0;0
16    Alchemilla vulgaris  <NA>  <NA>            <NA>         <NA>         <NA>      <NA>   <NA>   <NA>
17     Alliaria petiolata   120    20   2.25;3.4;3.03       5;5;5x        5;6;5     7;7;7  8;8;9  0;0;0
18       Allium angulosum  <NA>  <NA>            <NA>         <NA>         <NA>      <NA>   <NA>   <NA>
19     Allium lusitanicum  <NA>  <NA>            <NA>         <NA>         <NA>      <NA>   <NA>   <NA>
20       Allium oleraceum    80    25           0.881        6;7;7        3;5;4     7;7;7  4;4;4  0;0;0

I would be very grateful for any help!

Answer 1

You can write a function which splits the value on separator and calculates it's mean

split_and_mean <- function(x) {
   sapply(strsplit(x, ";"), function(x) mean(as.numeric(x)))
}

and then apply it on selected columns

df[4:9] <- lapply(df[4:9], split_and_mean)

df
#                       X h_max h_min seed_wght ell_light_uk ell_moist_uk ell_pH_uk ell_N ell_S
#1    Achilleamillefolium    45     8    0.1800         7.33         4.67        NA  4.67 1.000
#2       Achilleaptarmica    60    20    0.2550         7.33         7.67      4.33  3.00 0.000
#3   Aegopodiumpodagraria   100    40    2.0950           NA         5.67      6.67  7.67 0.333
#4     Agrimoniaeupatoria    60    30    7.9800         7.00         4.00      7.67  4.00 0.333
#5         Agrostiscanina    70    10    0.0550         7.67         8.33      3.00  2.33 0.000
#6     Agrostiscapillaris    70    10    0.0683         6.67           NA      4.00  4.00 0.000
#7       Agrostisgigantea    70    10    0.0850         7.00         6.67      6.67  6.33 0.333
#8    Agrostisstolonifera   100    15    0.0450         7.33           NA        NA  5.33 0.667
#9       Ajugachamaepitys    20     5    1.4101         7.33         3.67      8.33  2.67 0.000
#10       Ajugagenevensis  <NA>  <NA>        NA           NA           NA        NA    NA    NA
#11      Ajugapyramidalis    30    10    1.4865         7.00         5.00      2.33  1.67 0.000
#12          Ajugareptans    30    10    1.4000         5.33         6.33      5.67  5.67 0.000
#13      Alchemillaglabra    60  <NA>    0.5100           NA           NA        NA    NA    NA
#14 Alchemillaglaucescens    20  <NA>    0.4735         7.00         5.00      5.00  4.00 0.000
#15   Alchemillamonticola    40  <NA>    0.6890         6.33         4.67      6.00  4.00 0.000
#16    Alchemillavulgaris  <NA>  <NA>        NA           NA           NA        NA    NA    NA
#17     Alliariapetiolata   120    20    2.8933           NA         5.33      7.00  8.33 0.000
#18       Alliumangulosum  <NA>  <NA>        NA           NA           NA        NA    NA    NA
#19     Alliumlusitanicum  <NA>  <NA>        NA           NA           NA        NA    NA    NA
#20       Alliumoleraceum    80    25    0.8810         6.67         4.00      7.00  4.00 0.000

This returns warnings because there are columns with NA values that cannot be split but it is safe to ignore the warnings.

---

Similar concept using tidyverse , convert data to long format, split on ";", take mean and get data in wide format again.

library(tidyverse)
df %>%
  pivot_longer(cols = 4:9) %>%
   mutate(value = map_dbl(str_split(value, ";"), ~mean(as.numeric(.x)))) %>%
   pivot_wider()

data

Used this data to test the solutions. Make sure that you have column as characters and not factors before applying the solution.

df <- structure(list(X = c("Achilleamillefolium", "Achilleaptarmica", 
"Aegopodiumpodagraria", "Agrimoniaeupatoria", "Agrostiscanina", 
"Agrostiscapillaris", "Agrostisgigantea", "Agrostisstolonifera", 
"Ajugachamaepitys", "Ajugagenevensis", "Ajugapyramidalis", "Ajugareptans", 
"Alchemillaglabra", "Alchemillaglaucescens", "Alchemillamonticola", 
"Alchemillavulgaris", "Alliariapetiolata", "Alliumangulosum", 
"Alliumlusitanicum", "Alliumoleraceum"), h_max = c("45", "60", 
"100", "60", "70", "70", "70", "100", "20", "<NA>", "30", "30", 
"60", "20", "40", "<NA>", "120", "<NA>", "<NA>", "80"), h_min = c("8", 
"20", "40", "30", "10", "10", "10", "15", "5", "<NA>", "10", 
"10", "<NA>", "<NA>", "<NA>", "<NA>", "20", "<NA>", "<NA>", "25"
), seed_wght = c("0.16;0.2", "0.31;0.2", "1.99;2.2", "2.96;13", 
"0.05;0.06", "0.07;0.06;0.075", "0.08;0.09", "0.07;0.02", "1.6602;1.16", 
"<NA>", "1.373;1.6", "1.4", "0.51", "0.487;0.46", "0.689", "<NA>", 
"2.25;3.4;3.03", "<NA>", "<NA>", "0.881"), ell_light_uk = c("7;8;7", 
"7;8;7", "5;6x;6", "7;7;7", "7;7;9", "7;6;7", "7;7;7", "7;7;8", 
"7;7;8", "<NA>", "7;7;7", "5;5;6", "7;6;NA", "7;7;7", "6;7;6", 
"<NA>", "5;5;5x", "<NA>", "<NA>", "6;7;7"), ell_moist_uk = c("5;5;4", 
"8;7;8", "6;5;6", "4;4;4", "7;9;9", "4x;NA;5", "6;6;8", "6;7~;7", 
"4;3;4", "<NA>", "5;5;5", "6;7;6", "5;NA;6", "5;5;5", "5;4;5", 
"<NA>", "5;6;5", "<NA>", "<NA>", "3;5;4"), ell_pH_uk = c("6x;6;NA", 
"4;5;4", "6;7;7", "8;7;8", "3;3;3", "4;4;4", "6;7;7", "7;6x;NA", 
"8;9;8", "<NA>", "1;1;5", "6;5;6", "6;6;NA", "4;4;7", "6;6;6", 
"<NA>", "7;7;7", "<NA>", "<NA>", "7;7;7"), ell_N = c("4;5;5", 
"3;2;4", "8;8;7", "4;4;4", "2;3;2", "4;4;4", "6;6;7", "5;6;5", 
"2;2;4", "<NA>", "1;2;2", "5;6;6", "4;NA;6", "3;5;4", "4;4;4", 
"<NA>", "8;8;9", "<NA>", "<NA>", "4;4;4"), ell_S = c("1;1;1", 
"0;0;0", "0;0;1", "0;0;1", "0;0;0", "0;0;0", "0;1;0", "1;0;1", 
"0;0;0", "<NA>", "0;0;0", "0;0;0", "0;0;NA", "0;0;0", "0;0;0", 
"<NA>", "0;0;0", "<NA>", "<NA>", "0;0;0")), row.names = c("1", 
"2", "3", "4", "5", "6", "7", "8", "9", "10", "11", "12", "13", 
"14", "15", "16", "17", "18", "19", "20"), class = "data.frame")

Answer 2

With tidyverse , we can do

library(dplyr)
library(purrr)
df %>% 
   mutate_at(4:9, ~ strsplit(., ';') %>% 
                        map_dbl(~ mean(as.numeric(.x), na.rm = TRUE)))

data

df <- structure(list(X = c("Achilleamillefolium", "Achilleaptarmica", 
"Aegopodiumpodagraria", "Agrimoniaeupatoria", "Agrostiscanina", 
"Agrostiscapillaris", "Agrostisgigantea", "Agrostisstolonifera", 
"Ajugachamaepitys", "Ajugagenevensis", "Ajugapyramidalis", "Ajugareptans", 
"Alchemillaglabra", "Alchemillaglaucescens", "Alchemillamonticola", 
"Alchemillavulgaris", "Alliariapetiolata", "Alliumangulosum", 
"Alliumlusitanicum", "Alliumoleraceum"), h_max = c("45", "60", 
"100", "60", "70", "70", "70", "100", "20", "<NA>", "30", "30", 
"60", "20", "40", "<NA>", "120", "<NA>", "<NA>", "80"), h_min = c("8", 
"20", "40", "30", "10", "10", "10", "15", "5", "<NA>", "10", 
"10", "<NA>", "<NA>", "<NA>", "<NA>", "20", "<NA>", "<NA>", "25"
), seed_wght = c("0.16;0.2", "0.31;0.2", "1.99;2.2", "2.96;13", 
"0.05;0.06", "0.07;0.06;0.075", "0.08;0.09", "0.07;0.02", "1.6602;1.16", 
"<NA>", "1.373;1.6", "1.4", "0.51", "0.487;0.46", "0.689", "<NA>", 
"2.25;3.4;3.03", "<NA>", "<NA>", "0.881"), ell_light_uk = c("7;8;7", 
"7;8;7", "5;6x;6", "7;7;7", "7;7;9", "7;6;7", "7;7;7", "7;7;8", 
"7;7;8", "<NA>", "7;7;7", "5;5;6", "7;6;NA", "7;7;7", "6;7;6", 
"<NA>", "5;5;5x", "<NA>", "<NA>", "6;7;7"), ell_moist_uk = c("5;5;4", 
"8;7;8", "6;5;6", "4;4;4", "7;9;9", "4x;NA;5", "6;6;8", "6;7~;7", 
"4;3;4", "<NA>", "5;5;5", "6;7;6", "5;NA;6", "5;5;5", "5;4;5", 
"<NA>", "5;6;5", "<NA>", "<NA>", "3;5;4"), ell_pH_uk = c("6x;6;NA", 
"4;5;4", "6;7;7", "8;7;8", "3;3;3", "4;4;4", "6;7;7", "7;6x;NA", 
"8;9;8", "<NA>", "1;1;5", "6;5;6", "6;6;NA", "4;4;7", "6;6;6", 
"<NA>", "7;7;7", "<NA>", "<NA>", "7;7;7"), ell_N = c("4;5;5", 
"3;2;4", "8;8;7", "4;4;4", "2;3;2", "4;4;4", "6;6;7", "5;6;5", 
"2;2;4", "<NA>", "1;2;2", "5;6;6", "4;NA;6", "3;5;4", "4;4;4", 
"<NA>", "8;8;9", "<NA>", "<NA>", "4;4;4"), ell_S = c("1;1;1", 
"0;0;0", "0;0;1", "0;0;1", "0;0;0", "0;0;0", "0;1;0", "1;0;1", 
"0;0;0", "<NA>", "0;0;0", "0;0;0", "0;0;NA", "0;0;0", "0;0;0", 
"<NA>", "0;0;0", "<NA>", "<NA>", "0;0;0")), row.names = c("1", 
"2", "3", "4", "5", "6", "7", "8", "9", "10", "11", "12", "13", 
"14", "15", "16", "17", "18", "19", "20"), class = "data.frame")