CODEWITHSUNDEEP
rdplyrreshape
Nick Knauer
Nick Knauer

Reputation: 4243

Reformat Dataframe have Unique Date Rows

I have a dataframe below: 

dat

Day           Col1   Another    Value    High/Low
2018-01-01       A         B       20        High 
2018-01-01       B         G       30        High 
2018-01-02       C         D       40         Low
2018-01-02       C         M       70         Low
2018-01-02       C         G       50         Low

How do I reformat this so that it looks like this:

I want to order the reasons also in descending order if it's high and ascending if low. Also each date will never have multiple high/lows.

Day             Reason_1   Value_1    Reason_2    Value_2   Reason_3   Value_3  
2018-01-01           B-G        30         A-B         20         NA       NA
2018-01-02           C-D        40         C-G         50        C-M       70

I tried to use the spread function but that makes the Col1 and Another values as columns: 

library(tidyr)
new_df<- spread(dat, Col1_Another_Combined, Value)

No luck with this though

Upvotes: 0

Views: 37

Answers (1)

bouncyball
bouncyball

Reputation: 10761

We can use a variety of verbs from the tidyverse:

First, we group by day, and then calculate a ranking of the Value. We use row_number, because it will result in a unique ranking across the rows. We use unite to create the reason column, and then use gather, unite, and spread to do the final wide-to-long-to-wide transformation. One thing to note is that the value_* columns are still coded as character values, which can be easily changed.

dat %>%
    group_by(Day) %>%
    mutate(row_num_value = row_number(Value)) %>% # ranking
    unite(reason, Col1, Another, sep = " - ") %>% # reason column
    gather(variable, value, reason, Value) %>% # convert to long
    unite(variable2, variable, row_num_value, sep = '_') %>% 
    spread(variable2, value) # convert to wide

# A tibble: 2 x 8
# Groups:   Day [2]
  Day        High.Low reason_1 reason_2 reason_3 Value_1 Value_2 Value_3
  <chr>      <chr>    <chr>    <chr>    <chr>    <chr>   <chr>   <chr>  
1 2018-01-01 High     A - B    B - G    NA       20      30      NA     
2 2018-01-02 Low      C - D    C - G    C - M    40      50      70

Upvotes: 2

Related Questions