Repeating columns in reshape() or stack()

Question

I have this df:

df = data.frame(Meaning = c('Tax', 'Internet', 'Tax', 'Phone', 'Tax', 'Car'),
            Code = c(4656, 6152, 4656, 6150, 4656, 6151),
            Total = c(0.73, 4.4, 1.33, 8, 1.67, 10),
            Tax = c(0.73, NA, 1.33, NA, 1.67, NA),
            Subtotal = c(NA, 3.67, NA, 6.67, NA, 8.33),
             stringsAsFactors = FALSE)

> df
Meaning   Code   Total   Tax    Subtotal
Tax       4656   0.73    0.73   NA
Internet  6152   4.40    NA     3.67
Tax       4656   1.33    1.33   NA
Phone     6150   8.00    NA     6.67
Tax       4656   1.67    1.67   NA
Car       6151   10.00   NA     8.33

And I would like to use reshape() or stack to obtain another data.frame looking like this:

Code    Meaning   Category   Price
6152    Internet   Total      4.4
6152    Internet   Subtotal   3.67
4656    Tax        Subtotal   0.73
6150    Phone      Total      8
6150    Phone      Subtotal   6.67
4656    Tax        Subtotal   1.33
6151    Car        Total      10
6151    Car        Subtotal   8.33
4656    Tax        Subtotal   1.67

Where Category displays the column from df (Total or Subtotal) and Price is displayed in the following way: Total, Subtotal, Tax showed on df.

So far I tried with: cbind(df[1:2], stack(lapply(df[-c(1:2)], as.character)))

But it retrieves:

Meaning   Code values      ind
Tax       4656   0.73    Total
Internet  6152    4.4    Total
Tax       4656   1.33    Total
Phone     6150      8    Total
Tax       4656   1.67    Total
Car       6151     10    Total
Tax       4656   0.73      Tax
Internet  6152   <NA>      Tax
Tax       4656   1.33      Tax
Phone     6150   <NA>      Tax
Tax       4656   1.67      Tax
Car       6151   <NA>      Tax
Tax       4656   <NA> Subtotal
Internet  6152   3.67 Subtotal
Tax       4656   <NA> Subtotal
Phone     6150   6.67 Subtotal
Tax       4656   <NA> Subtotal
Car       6151   8.33 Subtotal

Any ideas?

NOTE: I have tried with all this answers but because of my df having some NA the solutions do not work. Answer 1, Answer 2, Answer 3

Why not use the new tidyverse package ``tidyr``. Here is a cheat sheet: [link](https://www.rstudio.com/resources/cheatsheets/). Use "Data Import Cheat Sheet", p. 2. The functions your looking for are ``spread`` and ``gather``. — xhr489, Feb 25 '19 at 13:18
Maybe the problem is that you should just filter the ``NA`` rows. — xhr489, Feb 25 '19 at 14:03

xhr489 · Answer 1 · 2019-02-25T13:37:12.627

Does this look right?

  library(tidyverse)

df %>% 
      gather(Total, Tax, Subtotal, key="key",  value="value") %>% 
      arrange(Code)

    Meaning Code      key value
1       Tax 4656    Total  0.73
2       Tax 4656    Total  1.33
3       Tax 4656    Total  1.67
4       Tax 4656      Tax  0.73
5       Tax 4656      Tax  1.33
6       Tax 4656      Tax  1.67
7       Tax 4656 Subtotal    NA
8       Tax 4656 Subtotal    NA
9       Tax 4656 Subtotal    NA
10    Phone 6150    Total  8.00
11    Phone 6150      Tax    NA
12    Phone 6150 Subtotal  6.67
13      Car 6151    Total 10.00
14      Car 6151      Tax    NA
15      Car 6151 Subtotal  8.33
16 Internet 6152    Total  4.40
17 Internet 6152      Tax    NA
18 Internet 6152 Subtotal  3.67

score 0 · Answer 2 · answered Feb 25 '19 at 13:57

I prefer to use the melt function from data.table:

library(data.table)

melt(df, 
     id.vars = c('Meaning', 'Code'), 
     variable.name = 'Category', 
     value.name = 'Price')

    Meaning Code Category price
1       Tax 4656    Total  0.73
2  Internet 6152    Total  4.40
3       Tax 4656    Total  1.33
4     Phone 6150    Total  8.00
5       Tax 4656    Total  1.67
6       Car 6151    Total 10.00
7       Tax 4656      Tax  0.73
8  Internet 6152      Tax    NA
9       Tax 4656      Tax  1.33
10    Phone 6150      Tax    NA
11      Tax 4656      Tax  1.67
12      Car 6151      Tax    NA
13      Tax 4656 Subtotal    NA
14 Internet 6152 Subtotal  3.67
15      Tax 4656 Subtotal    NA
16    Phone 6150 Subtotal  6.67
17      Tax 4656 Subtotal    NA
18      Car 6151 Subtotal  8.33

Repeating columns in reshape() or stack()

2 Answers2