Dzień dobry,
mam problem z zainstalowaniem pakietu BDL, w R Studio. Przy instalacji występuje błąd, ponieważ antywirus wykrywa zagrożenie i nie pozwala dokończyć instalacji pakietu. Co w tym wypadku należy zrobić, tak, żeby nie było potrzeby wyłączać antywirusa?

błąd_bdl.pdf

Under this SO question one can see the equivalent for plyr::join_all which is purrr:reduce.

> library(plyr)
> dfs <- list(
+   a = data.frame(x = 1:10, a = runif(10)),
+   b = data.frame(x = 1:10, b = runif(10)),
+   c = data.frame(x = 1:10, c = runif(10))
+ )
> join_all(dfs, "x")
    x          a         b         c
1   1 0.57119390 0.2653866 0.5130320
2   2 0.88501253 0.3914206 0.5990092
3   3 0.97230620 0.7255818 0.1687517
4   4 0.19460068 0.1252290 0.1923817
5   5 0.69917449 0.1464009 0.3980751
6   6 0.78650351 0.7987320 0.6369933
7   7 0.99910259 0.3822278 0.9299816
8   8 0.06227989 0.2037727 0.2606305
9   9 0.71419857 0.1844163 0.2498298
10 10 0.45508120 0.9791452 0.9281774
> 
> library(dplyr)
> purrr::reduce(dfs, left_join)
Joining, by = "x"
Joining, by = "x"
    x          a         b         c
1   1 0.57119390 0.2653866 0.5130320
2   2 0.88501253 0.3914206 0.5990092
3   3 0.97230620 0.7255818 0.1687517
4   4 0.19460068 0.1252290 0.1923817
5   5 0.69917449 0.1464009 0.3980751
6   6 0.78650351 0.7987320 0.6369933
7   7 0.99910259 0.3822278 0.9299816
8   8 0.06227989 0.2037727 0.2606305
9   9 0.71419857 0.1844163 0.2498298
10 10 0.45508120 0.9791452 0.9281774

I would recommend removing plyr from the package since it confuse the NAMESPACE with the dplyr package

> library(dplyr)

Attaching package: ‘dplyr’

The following objects are masked from ‘package:plyr’:

    arrange, count, desc, failwith, id, mutate, rename, summarise, summarize

and is not recommended even by the author.

Changes should be made in those places:

• https://github.com/statisticspoland/R_Package_to_API_BDL/blob/master/R/get_data_by_unit.R#L62
• https://github.com/statisticspoland/R_Package_to_API_BDL/blob/master/R/get_data_by_variable.R#L61
• https://github.com/statisticspoland/R_Package_to_API_BDL/blob/master/R/get_data_by_unit_locality.R#L72
• https://github.com/statisticspoland/R_Package_to_API_BDL/blob/master/R/get_data_by_variable_locality.R#L61

A common practice in R clients to APIs is to allow users to store keys in environment variables (e.g. here). It would be great for bdl to support that too. PR incoming.

If you'd like to reduce the dependenies amount in your package, you can consider removing packages that you import just to use one function. You could either copy the whole function into your package, or consider some substitution like nchar. nchar() doesn't handle NAs that properly as stringi_length()

> stringi::stri_length('')
[1] 0
> stringi::stri_length('NA')
[1] 2
> stringi::stri_length(NA)
[1] NA
> stringi::stri_length(NULL)
integer(0)
> stringi::stri_length('')
[1] 0
> 
> nchar('NA')
[1] 2
> nchar(NA)
[1] NA
> nchar(NULL)
integer(0)
> nchar('')
[1] 0

but you can always use

nchar_length <- function(x) {`if`(is.na(x), 0, nchar(x)) }; nchar_length(NA); nchar_length(''); nchar_length(''); nchar_length('asdasda')
[1] 0
[1] 0
[1] 0
[1] 7

instead of stringi::stri_length that imports the whole package and makes the user download the whole package during the installation

See this chapter about building and testing packages http://r-pkgs.had.co.nz/check.html#travis

To find out about the travis platform https://travis-ci.org/ that has some documentation for R workflow https://docs.travis-ci.com/user/languages/r/

CC @KaniaKrzysztof

https://kaniakrzysztof.github.io/bdl/ ->

Continuation on #7 https://travis-ci.org/statisticspoland/R_Package_to_API_BDL/builds/560501919

I can see it's listed in the .Rbuildignore https://github.com/statisticspoland/R_Package_to_API_BDL/blob/master/.Rbuildignore#L6

If you have that file locally, you should push it to the repository so one can also update the .Rmd the moment he wants to update the README.md. How you're gonna deal with the updates one made to README.md but to to README.Rmd

get_units() and search_units (search_units(name = "wro") etc. work well, however get and search subjects throws an error Status code:500

search_subjects("lud")
Error in get_request(dir, id, filters, ...) :
Failure to get data. Error occured. Status code: 500.

search_subjects(name = "lud")
Error in get_request(dir, id, filters, ...) :
Failure to get data. Error occured. Status code: 500.

search_subjects(name = "lud")
Error in get_request(dir, id, filters, ...) :
Failure to get data. Error occured. Status code: 500.

search_subjects(name = "lud", lang = "pl")
Error in get_request(dir, id, filters, ...) :
Failure to get data. Error occured. Status code: 500.

get_subjects()
Error in get_request(dir, id, filters, ...) :
Failure to get data. Error occured. Status code: 500.

get_subjects(parentId = "011212000000")
Error in get_request(dir, id, filters, ...) :
Failure to get data. Error occured. Status code: 500.

search_subjects("zdrowia")
Error in get_request(dir, id, filters, ...) :
Failure to get data. Error occured. Status code: 500.

get_subjects(Id = "011212000000")
Error in get_request(dir, id, filters, ...) :
Failure to get data. Error occured. Status code: 500.

get_units() and search_units (search_units(name = "wro") etc. work well.

Hi,

I'm trying to get the latest population for the statistical city, so I'm trying to use https://bdl.stat.gov.pl/api/v1/data/localities/by-unit/031613261011-0965016?var-id=60616 --> the latest population for Opole Instead I'm getting an empty list. Any idea what I'm doing wrong? @Nowosad @MarcinKosinski @KaniaKrzysztof @IwoA @statisticspoland

I have not found a way to understand how the "Territorial unit ID" identifies administrative regions, especially at the locality level where there are no shapefiles / SF geometries to locate them in space in the GIT.

Would be very useful to have a dictionary on how the Territorial Unit ID is constructed (to unite it with other sources such as https://geo.stat.gov.pl/en/inspire Or, even more useful, a (downloadable) SF file that has the same id so data from level 7 can be located in space.

Thanks for your time and sorry if I missed something.

On my Ubuntu 18 the instruction suggested in the README

remotes::install_github("statisticspoland/R_Package_to_API_BDL", upgrade = "always", type = "binary")

crashes with

Error in i.p(...) : type 'binary' is not supported on this platform

This one works

remotes::install_github("statisticspoland/R_Package_to_API_BDL")

In this line https://github.com/statisticspoland/R_Package_to_API_BDL/blob/master/DESCRIPTION#L17 one can see that roxygen2 is a suggested package. In this section user specify packages that are needed for a broader usage of the package: like vignettes' execution. None of users will ever need roxygen2 to use your package. This package is meant for developers, where the Suggests section is meant for users.

More http://r-pkgs.had.co.nz/description.html

Calling bdl::get_data_by_variable(54821, unitLevel = 6, year = 2011) I get the following error

Error in get_request(dir, id = attrId, filters, ...) : 
  Failure to get data. Probably invalid id. Status code: 412.

Traceback:

7: stop("Failure to get data. Probably invalid id. Status code: ", 
       status, msg)
6: get_request(dir, id = attrId, filters, ...)
5: attribute_info(attrId, lang = lang)
4: FUN(X[[i]], ...)
3: lapply(attributes, get_attr_label, lang = lang)
2: add_attribute_labels(df, lang)
1: bdl::get_data_by_variable(54821, unitLevel = 6, year = 2011)

The error occurs when requesting attributeInfo for attrId = 0. The request is made to https://bdl.stat.gov.pl/api/v1/attributes/0?lang=pl&page-Size=100&format=json which is invalid.

There is one row in the dataframe attributeIds are taken from with attrId = 0:

# A tibble: 1 x 5
  id           name                                  year    val attrId
  <chr>        <chr>                                 <chr> <int>  <int>
1 071412865998 GMINY-DZIELNICY WARSZAWY NIE USTALONO 2011      0      0

R_Package_to_API_BDL seems tought. Why not using some catchy name for the repository, that relates to the short name of the package? A repository could be named statisticspoland/rbld where the package, while still in the development phrase could also by renamed to rbld. Think about the other interface for other language that would probably can be named after laguages like
py-bdl and so on. With the rbld in the name, you not only have the catchy meanigful name for the repository but also you can distinguish between the bld package for R, Python and other languages.

Just a though, but it's better to solve this at the early beginning

If you got this https://github.com/statisticspoland/R_Package_to_API_BDL/blob/master/DESCRIPTION#L6
you can remove this https://github.com/statisticspoland/R_Package_to_API_BDL/blob/master/DESCRIPTION#L22

Package was archived on 2020-06-15 and can be no longer installed from CRAN. When we could expect a corrected version?

For now, I am not quite sure from which function I should start to build the whole workflow for the package. @KaniaKrzysztof can you provide a basic Vignette http://r-pkgs.had.co.nz/vignettes.html explaining the typical workflow, applications and the usage of the package?