diff --git a/README.Rmd b/README.Rmd index 96a7569..28276af 100644 --- a/README.Rmd +++ b/README.Rmd @@ -88,7 +88,7 @@ Otherwise the basic idea is: `devtools::install_github("tidyverse/readr#760")`. * For tables in html pages, use `unpivotr::tidy_html()` * For data frames, use `unpivotr::as_cells()` -- this should be a last - resort, because by the time thee data is in a conventional data frame, it + resort, because by the time the data is in a conventional data frame, it is often too late -- formatting has been lost, and most data types have been coerced to strings. 1. Either `behead()` straight away, else `dplyr::filter()` separately for the diff --git a/README.md b/README.md index 1759644..50e4509 100644 --- a/README.md +++ b/README.md @@ -86,7 +86,7 @@ Otherwise the basic idea is: `devtools::install_github("tidyverse/readr#760")`. - For tables in html pages, use `unpivotr::tidy_html()` - For data frames, use `unpivotr::as_cells()` – this should be a - last resort, because by the time thee data is in a conventional + last resort, because by the time the data is in a conventional data frame, it is often too late – formatting has been lost, and most data types have been coerced to strings. 2. Either `behead()` straight away, else `dplyr::filter()` separately @@ -99,6 +99,14 @@ Otherwise the basic idea is: ``` r library(unpivotr) library(tidyverse) +#> ── Attaching packages ────────────────────────────────── tidyverse 1.2.1 ── +#> ✔ ggplot2 3.1.0 ✔ purrr 0.2.5.9000 +#> ✔ tibble 1.4.99.9006 ✔ dplyr 0.7.8 +#> ✔ tidyr 0.8.2 ✔ stringr 1.3.1 +#> ✔ readr 1.2.1.9000 ✔ forcats 0.3.0 +#> ── Conflicts ───────────────────────────────────── tidyverse_conflicts() ── +#> ✖ dplyr::filter() masks stats::filter() +#> ✖ dplyr::lag() masks stats::lag() x <- purpose$`NNW WNW` x # A pivot table in a conventional data frame. Four levels of headers, in two #> X2 X3 X4 X5 X6 X7