Btukfyl

Is there a way that I can keep a separate list of headers that basically acts like a dictionary that lists the descriptive header and then an easier to use short name for each header that I could call back and forth without needing to maintain the correct order of the columns? I'm not great with this but here is an example of what I was thinking:

Original Data Set

|---------------------|------------------|------------------|

|    Descriptive A    |  Descriptive B   |  Descriptive C   |

|---------------------|------------------|------------------|

|          12         |         34       |       25         |

|---------------------|------------------|------------------|

Dictionary of Headers

|---------------------|------------------|

|      long_name      |    short_name    |

|---------------------|------------------|

|      Descriptive A  |         A        | 

|---------------------|------------------|

|      Descriptive B  |         B        | 

|---------------------|------------------|

|      Descriptive C  |         C        | 

|---------------------|------------------|

Then I could have a piece of code that calls on the short_name column of the dictionary to replace the long_name title of the headers with the short_name and then I would not have to rely on the position of headers.

I'm not sure if that is possible but I have a table with 180 columns (that's growing) and they all have descriptive names that don't translate well into R, so I thought this might be a solution that I could continue to add to as the data set grows.

Yes, you just need the dictionary (or codebook) as a separate data frame (can be read in from, say, a .csv file). Let's say you have a data frame like this:

df <- data.frame(matrix(rnorm(1000), ncol = 100))

names(df) <- paste0("a very long unfortunate name to be replaced_", 1:ncol(df))

You can create the codebook like this:

codebook <- data.frame(long_name = names(df), short_name = paste0("X_", 1:ncol(df)),

                       stringsAsFactors = F)



                                      long_name short_name

1 a very long unfortunate name to be replaced_1        X_1

2 a very long unfortunate name to be replaced_2        X_2

3 a very long unfortunate name to be replaced_3        X_3

4 a very long unfortunate name to be replaced_4        X_4

5 a very long unfortunate name to be replaced_5        X_5

6 a very long unfortunate name to be replaced_6        X_6

Let's then change names of df using the "short names"

names(df) <- codebook[ ,2]

For fun, let's randomise the rows of codebook to show you cna still use it:

codebook <- codebook[sample(nrow(codebook)), ]

Finally you can use match() to retrieve the original long names:

codebook$long_name[match(names(df), codebook$short_name)]



  [1] a very long unfortunate name to be replaced_1   a very long unfortunate name to be replaced_2  

  [3] a very long unfortunate name to be replaced_3   a very long unfortunate name to be replaced_4  

  [5] a very long unfortunate name to be replaced_5   a very long unfortunate name to be replaced_6  

  [7] a very long unfortunate name to be replaced_7   a very long unfortunate name to be replaced_8  

  [9] a very long unfortunate name to be replaced_9   a very long unfortunate name to be replaced_10

Like I commented, I don't think there's a way to do aliasing in place, but for calculation you can do something like:

df1 <- data.frame(

  "Descriptive A" = 12,

  "Descriptive B" = 34,

  "Descriptive C" = 25,

  check.names = FALSE

)

The "aliasing" object can be a frame, but since all you're doing is assigning a name to a name, it is efficiently handled by a named character vector:

df1_aliases <- c(

  "B" = "Descriptive B",

  "A" = "Descriptive A",

  "C" = "Descriptive C"

)

Your aliases steps would be an intentional pre-/post-translation of names:

names(df1) <- names(df1_aliases)[ match(names(df1), df1_aliases) ]

df1

#    A  B  C

# 1 12 34 25



### do stuff here ###



names(df1) <- df1_aliases[ match(names(df1), names(df1_aliases)) ]

df1

#   Descriptive A Descriptive B Descriptive C

# 1            12            34            25

It might be feasible to overwrite $.data.frame and $<-.data.frame for basic dollar-sign operations, but you'd also need to overwrite [.data.frame, [[.data.frame, and perhaps even with (depending on your frame-access habits) ... and those rewritten functions might not work from all other functions you are using (depending on their function/namespace search path).

Because of the complexities of tracking down everything that touches the frame, I strongly suggest you make it as explicit as possible: have only one set of names each column is known by (whether the original or your aliases), never both simultaneously. This means the translate/untranslate steps are explicit and anything that works on the frame will work unambiguously.

You could give the names names, and then subset the names before subsetting the data.frame.

For example, using the iris data:

short_names <- names(iris)

names(short_names) <- c("sl","sw","pl","pw","sp")

attributes(iris)$names <- short_names



head(iris[names(iris)[c("sl","sp")]])

  Sepal.Length Species

1          5.1  setosa

2          4.9  setosa

3          4.7  setosa

4          4.6  setosa

5          5.0  setosa

6          5.4  setosa

Using dict and DF defined reproducibly in the Note at the end run the for loop
shown and then we can use A, B and C without quotes as column names.

for(i in 1:nrow(dict)) assign(dict$short_name[i], dict$long_name[i])



# test - use DF[B] in place of DF["Descriptive B"]

DF[B]

##   Descriptive B

## 1            34

As shown in the above test it is straight forward when using conventional subscripting. If you want to use nonstandard evaluation such as in dplyr then you will need to use rlang in the usual way:

library(dplyr)



DF %>% mutate(D = !!sym(B))

##   Descriptive A Descriptive B Descriptive C  D

## 1            12            34            25 34

Note

We assume this input:

Lines1 <- "

      long_name      |    short_name

      Descriptive A  |         A     

      Descriptive B  |         B     

      Descriptive C  |         C" 

dict <- read.table(text = Lines1,  header = TRUE, sep = "|", as.is = TRUE, 

  strip.white = TRUE)



Lines2 <- "

    Descriptive A    |  Descriptive B   |  Descriptive C 

          12         |         34       |       25"

DF <- read.table(text = Lines2,  header = TRUE, sep = "|", as.is = TRUE,

   strip.white = TRUE, check.names = FALSE)