Given a dataframe df as follows:
df <- structure(list(code = c("M0000273", "M0000357", "M0000545", "M0000273",
"M0000357", "M0000545"), name = c("industry", "agriculture",
"service", "industry", "agriculture", "service"), act_value = c(16.78,
9.26, 49.38, 35.74, 88.42, 68.26), pred_value = c(17.78, 10.26,
50.38, 36.74, 89.42, 69.26), year = c(2019L, 2019L, 2019L, 2020L,
2020L, 2020L)), class = "data.frame", row.names = c(NA, -6L))
df:
code name act_value pred_value year
1 M0000273 industry 16.78 17.78 2019
2 M0000357 agriculture 9.26 10.26 2019
3 M0000545 service 49.38 50.38 2019
4 M0000273 industry 35.74 36.74 2020
5 M0000357 agriculture 88.42 89.42 2020
6 M0000545 service 68.26 69.26 2020
I would like to use code and name as index columns, and convert act_value and pred_value from long to wide, and finally rename new columns by paste year column as prefix.
The expected result will like to the format as follows:
code name 2019_act_value 2019_pred_value 2020_act_value 2020_pred_value
1 M0000273 industry 16.78 17.78 35.74 36.74
2 M0000357 agriculture 9.26 10.26 88.42 89.42
3 M0000545 service 49.38 50.38 68.26 69.26
My trial code:
reshape(df, idvar = c('code', 'name'), timevar = 'year', direction = 'wide')
How could I achieve that correctly using R? Thanks.
>Solution :
We can use tidyr::pivot_wider to do this. I wouldn’t recommend your naming convention, and if you drop names_glue we get the same result but with the tidier year as suffix format instead.
library(tidyr)
pivot_wider(df,
names_from = year,
names_glue = "{year}_{.value}",
values_from = ends_with("value"))
#> # A tibble: 3 × 6
#> code name `2019_act_value` `2020_act_value` `2019_pred_value`
#> <chr> <chr> <dbl> <dbl> <dbl>
#> 1 M0000273 industry 16.8 35.7 17.8
#> 2 M0000357 agriculture 9.26 88.4 10.3
#> 3 M0000545 service 49.4 68.3 50.4
#> # … with 1 more variable: 2020_pred_value <dbl>