r reshape2 melt returns Warning in if (drop.margins) { : the condition has length > 1 and only the first element will be used

Question

I have a piece of code that returns a warning message:

Warning in if (drop.margins) { :
  the condition has length > 1 and only the first element will be used

from deep inside the reshape2 melt function as shown in the error message.

How do I correct?

The code is difficult to subset so I've included a description of the data frame. I'm just looking for a hint.

S_Melted <- melt(S_Flattened, S_Flattened$db_date)

BTW: S_Flattened was created by a cast in an earlier statement:

S_Flattened = cast(S, db_date ~ MetricType, value="AvgValue")

Answer 1

There are a couple of problems here:

1. You're actually using the "reshape" package, not "reshape2".
2. You're specifying a vector of values as the "id" variable, hence this warning.

Consider the following:

long <- structure(list(ID = structure(c(1L, 2L, 3L, 1L, 2L, 3L), .Label = c("A", 
    "B", "C"), class = "factor"), variable = structure(c(1L, 1L, 
    1L, 2L, 2L, 2L), .Label = c("V1", "V2"), class = "factor"), value = 1:6), 
    .Names = c("ID", "variable", "value"), row.names = c(NA, 6L), class = "data.frame")

Using cast from "reshape" gives you something that looks like a data.frame, but which has a bunch of other attributes.

reshape_df <- reshape::cast(long, ID ~ variable)
str(reshape_df)
# List of 3
#  $ ID: Factor w/ 3 levels "A","B","C": 1 2 3
#  $ V1: int [1:3] 1 2 3
#  $ V2: int [1:3] 4 5 6
#  - attr(*, "row.names")= int [1:3] 1 2 3
#  - attr(*, "idvars")= chr "ID"
#  - attr(*, "rdimnames")=List of 2
#   ..$ :'data.frame':  3 obs. of  1 variable:
#   .. ..$ ID: Factor w/ 3 levels "A","B","C": 1 2 3
#   ..$ :'data.frame':  2 obs. of  1 variable:
#   .. ..$ variable: Factor w/ 2 levels "V1","V2": 1 2

Here's your warning:

reshape::melt(reshape_df, reshape_df$ID)
#      ID value variable
# V1    A     1       V1
# V1.1  B     2       V1
# V1.2  C     3       V1
# V2    A     4       V2
# V2.1  B     5       V2
# V2.2  C     6       V2
# Warning message:
# In if (drop.margins) { :
#   the condition has length > 1 and only the first element will be used

And the same thing, without a warning.

reshape::melt(reshape_df, id = "ID")
#      ID value variable
# V1    A     1       V1
# V1.1  B     2       V1
# V1.2  C     3       V1
# V2    A     4       V2
# V2.1  B     5       V2
# V2.2  C     6       V2

---

A better approach would be to stop using "reshape" and start using "reshape2", "data.table" (which provides more flexible implementations of melt and dcast than "reshape2" does), or "tidyr".

Here's the same set of steps with "reshape2":

reshape2_df <- reshape2::dcast(long, ID ~ variable)

You get back a standard data.frame with no extra attributes.

str(reshape2_df)
# 'data.frame': 3 obs. of  3 variables:
#  $ ID: Factor w/ 3 levels "A","B","C": 1 2 3
#  $ V1: int  1 2 3
#  $ V2: int  4 5 6

melting is not a problem either -- just don't supply it with a vector, as you did in your attempt.

reshape2::melt(reshape2_df, "ID")
#   ID variable value
# 1  A       V1     1
# 2  B       V1     2
# 3  C       V1     3
# 4  A       V2     4
# 5  B       V2     5
# 6  C       V2     6