svraka/asmisc

🧰 Miscellaneous R utility functions

asmisc's Issues

Write tests

In GitLab by @svraka on Feb 26, 2019, 14:34

• Column order and contents in codebook with and without character/numeric/date variables (including chr_values)
• Default skims are restored even after failed codebook()
• codebook_chunked attributes are added

Prepare for skimr v2

In GitLab by @svraka on Feb 27, 2019, 15:04

Major changes coming, see skimr-v2-roadmap.

Store parsing problems as a tibble in codebook_chunked()

In GitLab by @svraka on Mar 22, 2019, 10:53

Return column specification in `read_delim_chunked_to_dataset()`

readr::read_delim_chunked writes the – abbreviated – spec to console but it would be nice to return the whole spec. Could be an attribute of the return value?

Show values in chr_values when the column is all numeric

In GitLab by @svraka on Mar 22, 2019, 14:15

Print warnings on parsing errors with `parse_date_ymd()`

It uses lubridate:.parse_date_time2() in the background, which unlike lubridate:.parse_date_time2() doesn't omit any warnings on parsing errors. The change would require some wrangling with centuries to keep the current function signature.

Add column on 32/64 bit integer

In GitLab by @svraka on Mar 22, 2019, 09:19

read_chunked() hides some readr options

In GitLab by @svraka on Mar 21, 2019, 15:36

read_chuned() should be more explicit about readr options. E.g. input_delim = ";" chooses read_csv2, which will use , as decimal and . as grouping mark. This could lead to unfortunate side effects. A better solution would be to use read_delim() and set the options manually. However, some of the niceness of read_csv(), like automatic whitespace remove should be kept.

Use a more robust solution in `find_project_packages()`

My regex hack fails often. attachment also seems to use regex based approach is better tested.