PowerQuery Puzzle solved with R

#223–224

Puzzles

Author: ExcelBI

All files (xlsx with puzzle and R with solution) for each and every puzzle are available on my Github. Enjoy.

Puzzle #223

As usual on weekends we are mainly doing table transformations. Sometimes it need simple manouvers and sometimes very complicated almost magical tricks. And today we have exactly that situation. First one is like: squeeze, pull, push, twist. Simple usage of basic functions.

If you are curious, why I illustrated it with scene of cafeteria… We sometimes need to complete tables like lunch on tray. :D

Loading libraries and data

library(tidyverse)
library(readxl)

path = "Power Query/PQ_Challenge_223.xlsx"
input = read_excel(path, range = "A1:D14")
test  = read_excel(path, range = "F1:J8")

Transformation

result = input %>%
  unite("Code", c("Type", "Code"), sep = "") %>%
  mutate(col = ifelse(row_number() %% 2 == 0, 2, 1), 
         row = (row_number() + 1) %/% 2,
         .by = Group) %>%
  pivot_wider(names_from = col, values_from = c(Code, Value), names_sep = "") %>%
  select(-row)

Validation

all.equal(result, test)
#> [1] TRUE

Puzzle #224

And the second one is one of the most complicated to solve because it is totally untidy. We have different data in the same row, we have headers every couple of rows. Fortunatelly I found a way to rebuild it in less than 15 LOC.

Loading libraries and data

library(tidyverse)
library(readxl)
library(janitor)

path = "Power Query/PQ_Challenge_224.xlsx"
input = read_excel(path, range = "A1:D12")
test  = read_excel(path, range = "F1:I20")

Transformation

result = input %>%
  mutate(date = ifelse(str_detect(Column1, "\\d"), Column1, NA)) %>%
  fill(date) %>%
  set_names(.[1, ]) %>%
  rename("Name" = 1, "date" = 5) %>%
  filter(!str_detect(Name, "\\d")) %>%
  mutate(date = coalesce(excel_numeric_to_date(as.numeric(date)), mdy(date))) %>%
  pivot_longer(-c(date, Name), names_to = "Data", values_to = "Value") %>%
  na.omit() %>%
  select(Date = date, Name, Data, Value) %>%
  mutate(Value = as.numeric(Value), 
         Date = as.POSIXct(Date))

Validation

all.equal(result, test, check.attributes = FALSE)
#> [1] TRUE

Feel free to comment, share and contact me with advices, questions and your ideas how to improve anything. Contact me on Linkedin if you wish as well.