Today, PX-Win has taken over the offline handling of PX files from PC-Axis and shares a codebase with PXWeb, but still remains somewhat in the shadows.
The lack of showcases for PX-Win limits its potential use in operations.
Often, the producer of a PX file does not have direct access to a server with PXWeb and, therefore, cannot fully benefit from PXWeb as cube explorer. A dataset does not need extensive metadata for PXWin to be useful, but naturally, more detail is required before it can be used for dissemination on PXWeb.
Step 1: Prepare data
In this case, simply get selected columns from the diamond dataset in ggplot2 (part of tidyverse)
next add these magic functions in one line of code:
count(across(everything()))
library(tidyverse)
diamonds <- ggplot2::diamonds %>%
select(carat,cut,color,clarity) %>%
count(across(everything()))
head(diamonds)# A tibble: 6 x 5
carat cut color clarity n
<dbl> <ord> <ord> <ord> <int>
1 0.2 Very Good E VS2 1
2 0.2 Premium D VS2 2
3 0.2 Premium E SI2 1
4 0.2 Premium E VS2 4
5 0.2 Premium F VS2 1
6 0.2 Ideal D VS2 1Step 2: add pxmake
use 2 functions from pxmake:
px() - Create a px object from a px file, an Excel metadata workbook, a data frame or a Parquet file.
px_save - Save px object to file
for help, check vignettes ’?px’and open with Pxwin.
# install.packages('devtools')
# devtools::install_github('StatisticsGreenland/pxmake')
library(pxmake)
px(diamonds) %>%
px_save("explore1.px")
Step 2: add totals
before saving, add totals
px(diamonds) %>%
px_add_totals(c("carat","cut","color","clarity")) %>%
px_save("explore2.px")
Step 3: and elimination
eliminate with ‘Total’
px(diamonds) %>%
px_elimination('Total') %>%
px_add_totals(c("carat","cut","color","clarity")) %>%
px_save("explore3.px")
Step 4: play
Go ahead, get PXWin for windows, unpack and play:
https://www.scb.se/en/services/statistical-programs-for-px-files/install-px-win/
That’s what they usually say, and then I tell them—go work your magic somewhere else, save it as a Parquet file, and load it into pxmake.
# install.packages('arrow')
library(arrow)
write_parquet(diamonds,"magic.parquet")
px("magic.parquet") %>%
px_elimination('Total') %>%
px_add_totals(c("carat","cut","color","clarity")) %>%
px_save("magic.px")