Can you add some parameters?
Closed this issue · 3 comments
This package is indeed very good. Recently, it has solved my problem of reading large xlsx, but I hope the author can add more custom parameters, such as
- Specify the data type for each column,
- Specify the na value
- When reading data, do not introduce Scientific notation, especially if there are both text and numbers in a column, text will be selected by default, but numbers will be recognized as scientific counting
- There seems to be a coding issue? (test file)
> mm = SheetReader::read_xlsx(path = f2, sheet = 1)
> head(mm$Profit)
[1] "本期利润" "没有单位"
[3] NA "1.72925e+08"
[5] NA NA
Thank you,
I have pushed a fix for 4., there was an issue with xml-escaped unicode characters. If you have devtools
you can try to install via install_github("fhenz/SheetReader-r")
, I will probably only upload a new CRAN version once I have also addressed some of your other points.
I think 1. and 2. are both good ideas, I will try to implement something similar to what readxl
also has.
3. is a bit tricky because Excel doesn't differentiate between integer or real numbers when storing, but I should be able to solve this more elegantly if I implement 1. (so it would then be solved by specifiying string/text as the column data type, that should be sufficient?).
Thank you, I have pushed a fix for 4., there was an issue with xml-escaped unicode characters. If you have
devtools
you can try to install viainstall_github("fhenz/SheetReader-r")
, I will probably only upload a new CRAN version once I have also addressed some of your other points.I think 1. and 2. are both good ideas, I will try to implement something similar to what
readxl
also has. 3. is a bit tricky because Excel doesn't differentiate between integer or real numbers when storing, but I should be able to solve this more elegantly if I implement 1. (so it would then be solved by specifiying string/text as the column data type, that should be sufficient?).
Thank you for your reply. Indeed, if 1 is resolved, then 3 can theoretically be resolved,
A new parameter col_types
has been added that allows specifying the data types for columns via named/unnamed character vector, e.g. read_xlsx([...], col_types=c("Profit"="text"))
.