How can I create a summary table from lists of matrices?

Question

Data:

k1 <- 
structure(list(`3280.2000` = structure(c(0, 1.363, 0, 4.198, 
1.097, 1.669, 0, 1.098, 0), .Dim = c(3L, 3L), .Dimnames = list(
    c("3280.3500", "3280.400", "3280.4000"), c("3280.3500", "3280.400", 
    "3280.4000"))), `160.2000` = structure(c(0.935, 0.929, 1.1, 
1.152), .Dim = c(2L, 2L), .Dimnames = list(
    c("160.2200", "160.2300"), c("160.2200", "160.2300")))), .Names = c("3280.2000", "160.2000"))

#k1 Looks like
# $`3280.2000`
#              3280.3500 3280.400 3280.4000
# 3280.3500     0.000    4.198     0.000
# 3280.400      1.363    1.097     1.098
# 3280.4000     0.000    1.669     0.000
# 
# $`160.2000`
#           160.2200 160.2300
# 160.2200    0.935    1.100
# 160.2300    0.929    1.152

k2 <- 
structure(list(`3280.2000` = structure(c(0, 1.363, 0, 4.198, 
1.097, 1.669, 0, 1.098, 0), .Dim = c(3L, 3L), .Dimnames = list(
    c("400","3500",  "4000"), c("400", "3500",  
    "4000"))), `160.2000` = structure(c(0.935, 0.929, 1.152, 
0.903), .Dim = c(2L, 2L), .Dimnames = list(c("2200", "2300"
), c("2200", "2300")))), .Names = c("3280.2000", "160.2000"
))

#k2 Looks like
# $`3280.2000`
#      400  3500  4000
# 400  0.000 4.198 0.000
# 3500 1.363 1.097 1.098
# 4000 0.000 1.669 0.000
# 
# $`160.2000`
#      2200  2300
# 2200 0.935 1.152
# 2300 0.929 0.903

Situation: I have a list of matrices that look as described above. I have matrices of different sizes e.g. (3x3, 2x2, etc) in lists. Each list and each matrix is perfectly similar except for the naming and sorting. I have k=2 (k1 and k2 above) lists I am trying to merge together in a summary table.

Objective: Retrieve data and names from each list, then merge to create summary table.

Possible solutions: a) Since the matrices are different sizes it makes the most sense to reshape each matrix in each list from square to column. Melt.list() is a good candidate. b) Change names of rows & columns in matrix. Since the 3280 is in the name of the object, it doesn't need to be in the name of the matrix row/columns. Thus row and column could provide a unique ID which could be successfully melted.

Problems:

a) Melt (or melt.list) truncates my variable names. This upsets me (X1 and X2 below should be 7-8 digits).

#   X1      X2 value        L1
# 1  3280.35 3500.00 0.000 3280.2000
# 2  3280.40 3500.00 1.363 3280.2000
# 3  3280.40 3500.00 0.000 3280.2000
# etc

b) I can change the colnames for a matrix, but I can't figure out how to do it for multiple matrices in an array. When I try this, I get a list of names (no more values!) rather than an update to the names.

#this works
colnames(k1[[1]]) <- gsub("*.*\\.", "", colnames(k1[[1]]))

#this does not - returns names where values should be!
k1 <- lapply(k1, function (x) {colnames(x) <- gsub("*.*\\.", "", colnames(x))}

Question: How can I either: a) prevent melt from truncating my names or b) change dimnames on a list of matrices?

Lists, names, frames, strata: how many ways can I fix my data?

Searched: reshape2 documentation, plyr documentation, [r] rename matrix columns; column renaming, changing row and column names, how to reshape dataframe, and plenty of others

Thanks in advance.

Ricardo Saporta Ricardo Saporta · Accepted Answer · 2013-04-07T19:51:28

The reason your lapply statement ins't giving you the desired output is because you forgot to return x.

# instead of: 
h <- lapply(h, function (x) {colnames(x) <- gsub("*.*\\.", "", colnames(x))})

# Try this: 
h <- lapply(h, function (x) {colnames(x) <- gsub("*.*\\.", "", colnames(x)); x})

How can I create a summary table from lists of matrices?

1 Answers