filtering rows based on multiple conditions-CodePudding

With the following dataframe, I would like to filter out rows based on two conditions. If the daily temperature is less than 41 I would like to extract the epi temperature, if it is >= 41 I would like to extract the meta temperature.

df <- data.frame(day = c(1, 1, 1, 2, 2, 2),
           temperature = c(40, 39, 39, 45, 38, 30),
           strata = c("epi", "meta", "hypo", "epi", "meta", "hypo"))

desired output:

day temperature strata
1    40         epi
2    38         meta

I cant quite wrap my head around how to code this using tidyverse.

CodePudding user response：

Here is a base R solution with ave. Create a logical index on the temperature and keep the rows where the strata condition is met, conditional to the created index.

df <- data.frame(day = c(1, 1, 1, 2, 2, 2),
                 temperature = c(40, 39, 39, 45, 38, 30),
                 strata = c("epi", "meta", "hypo", "epi", "meta", "hypo"))

i <- as.logical(with(df, ave(temperature, day, FUN = \(x) all(x < 41))))
df[(i & df$strata == "epi") | (!i & df$strata == "meta"), ]
#>   day temperature strata
#> 1   1          40    epi
#> 5   2          38   meta

^{Created on 2022-07-21 by the reprex package (v2.0.1)}

And here is a dplyr solution.

suppressPackageStartupMessages(library(dplyr))

df %>%
  group_by(day) %>%
  filter(
    (any(temperature >= 41) & strata == "meta") |
      (all(temperature < 41) & strata == "epi")
  )
#> # A tibble: 2 × 3
#> # Groups:   day [2]
#>     day temperature strata
#>   <dbl>       <dbl> <chr> 
#> 1     1          40 epi   
#> 2     2          38 meta

^{Created on 2022-07-21 by the reprex package (v2.0.1)}