Split function not maintaining structure of dataframe? - r

Question

I am doing hierarchical clustering in R and need all the cluster's elements separately.
When I use following data splits into 3 list of num [1:2628] (no info of columns in original dataframe (dataA) is transferred)
clusterA <- hclust(dist(dataA),method = "single")
NumA = 3
label <- cutree(clusterA, NumA)
clusterXlist<-split(dataA,f=label)
str(clusterXlist[[1]])
how to make shure that it maintains the structure of dataA
edit:
in my case
>str(clusterXlist[[1]])
num [1:2628] 0.0529 -0.3909 -0.4465 0.1 0.8393 ...
where as for dataA
> str(dataA)
num [1:440, 1:6] 0.0529 -0.3909 -0.4465 0.1 0.8393 ...
- attr(*, "dimnames")=List of 2
..$ : NULL
..$ : chr [1:6] "Fresh" "Milk" "Grocery" "Frozen" ...
- attr(*, "scaled:center")= Named num [1:6] 12000 5796 7951 3072 2881 ...
..- attr(*, "names")= chr [1:6] "Fresh" "Milk" "Grocery" "Frozen" ...
- attr(*, "scaled:scale")= Named num [1:6] 12647 7380 9503 4855 4768 ...
..- attr(*, "names")= chr [1:6] "Fresh" "Milk" "Grocery" "Frozen" ...
edit2 :
for dataA
> dput(head(dataA,n=20))
structure(c(0.0528730042415329, -0.390857056063646, -0.44652098379972,
0.0999975794271863, 0.839284119671916, -0.204572661537808, 0.00993903725191922,
-0.349583518736614, -0.477357534676238, -0.473957607271904, -0.682697336282181,
0.0905884780058897, 1.55872457204484, 0.728746944991474, 1.00042486502152,
-0.138155475034538, -0.868191050016313, -0.484236457564077, 0.521904849881291,
-0.333690834823332, 0.522972471408079, 0.543838613660349, 0.408073194590386,
-0.623310408164662, -0.0523368792616442, 0.333686752405346, -0.351915064454946,
-0.113851350576777, -0.291078065290861, 0.717677967619194, -0.053285340273111,
-0.63306600713975, 0.883794139056095, 0.0557876760455718, 0.497093035238056,
-0.634420951441845, 0.409157150032062, 0.0488774601048851, 0.0719115132405076,
-0.447303143322465, -0.0410681453901357, 0.170124700204028, -0.0281250860936324,
-0.3925300807586, -0.0792659545334748, -0.297298628211157, -0.10273182626616,
0.15518230654465, -0.185125447641461, 1.15011422238562, 0.528531691780372,
-0.360751187201331, 0.400469064432042, 0.739829765498898, 0.435615257968889,
-0.434621330503326, 0.438772101699743, -0.528063904936618, 0.226000834240152,
0.159180975270399, -0.588697039406295, -0.269829034507317, -0.137379339965946,
0.68636300602308, 0.173661155768845, -0.495590877769126, -0.533904475256987,
-0.288985833251248, -0.545233764836731, -0.394039245717966, 0.273564891153861,
-0.340276616984998, -0.573659982327726, 0.00475174748902491,
-0.572218072744849, -0.551001403168238, -0.605176006067741, -0.459955112363749,
-0.178576756619561, -0.494972916519322, -0.0435191938188023,
0.0863085949200282, 0.13308015693741, -0.498021323377842, -0.23165413161966,
-0.227878848586867, 0.0542186891412866, 0.0921812574154842, -0.244448146341904,
0.952945788892319, 0.649245242698738, -0.489212329634658, 0.209634507324604,
0.802353943473126, 0.456496070080021, -0.40217108193415, 0.341140199633565,
-0.526755422016323, -0.0240135648160378, -0.0762383134363428,
-0.066263629344282, 0.0890496850231094, 2.24074190324533, 0.0933048443208461,
1.29786952218849, -0.0261942126239276, -0.347458739603052, 0.369181005457445,
-0.274766434933383, 0.203229792845712, 0.0777025935624781, -0.364479376793999,
0.498608767430271, -0.327246732938803, 0.228051555415843, -0.394620088486301,
-0.157749554245622, 1.04716972023017, 0.587257919466454, -0.36306099036142
), .Dim = c(20L, 6L), .Dimnames = list(NULL, c("Fresh", "Milk",
"Grocery", "Frozen", "Detergents_Paper", "Delicassen")))
for clusterXlist[[1]] which was obtained by split of dataA
> dput(head(clusterXlist[[1]],n=20))
c(0.0528730042415329, -0.390857056063646, -0.44652098379972,
0.0999975794271863, 0.839284119671916, -0.204572661537808, 0.00993903725191922,
-0.349583518736614, -0.477357534676238, -0.473957607271904, -0.682697336282181,
0.0905884780058897, 1.55872457204484, 0.728746944991474, 1.00042486502152,
-0.138155475034538, -0.868191050016313, -0.484236457564077, 0.521904849881291,
-0.333690834823332)

Answer 1

What you have there is a matrix, not a data frame.
class(dataA)
# [1] "matrix"
The quick and easy way to split() would be to do
split(as.data.frame(dataA), label)
However, this may cause issues in later calculations and you may need to resort to coercing those list elements back to a matrix. I would recommend you use lapply() to split the data, as follows.
clusterXlist <- lapply(
unique(label),
function(i) dataA[label == i, , drop = FALSE]
)
to properly maintain your matrix structure throughout your list elements.
str(clusterXlist[[1]])
# num [1:18, 1:6] 0.0529 -0.3909 0.1 0.8393 -0.2046 ...
# - attr(*, "dimnames")=List of 2
# ..$ : NULL
# ..$ : chr [1:6] "Fresh" "Milk" "Grocery" "Frozen" ...