Passing along from readxl tidyverse/readxl#564.

User has 6 xls files, one of which can be read by readxl and xls2csv. The remaining five do not throw an error, but they produce no output.

I note I can read these files with a completely separate tool (the R package gdata, which wraps a Perl script 😬), but only if I specify the encoding: e.g., as read.xls(files[2], fileEncoding="latin1").

I've attempted to pass encoding to xls2csv but it does not change my results. That could be user error because I've never really specified encoding to libxls before.

I can also confirm what the user reports: merely opening and closing the problematic files in Excel makes them readable by readxl/libxls. I suppose this is changing the encoding?

The attached ZIP file is the one provided by my user in the readxl issue. I contains 6 directories, each of which holds one .txt and one .xls file.

fs::dir_tree("investigations/Data")
#> investigations/Data
#> ├── 20190326.seq
#> │   ├── 2019-3-26-.txt
#> │   └── 33.0000.XLS
#> ├── 20190327.seq
#> │   ├── 01.0000.XLS
#> │   └── 2019-3-27-.txt
#> ├── 20190328.seq
#> │   ├── 04.0000.XLS
#> │   └── 2019-3-28-.txt
#> ├── 20190329.seq
#> │   ├── 15.0000.XLS
#> │   └── 2019-3-29-.txt
#> ├── 20190330.seq
#> │   ├── 09.0000.XLS
#> │   └── 2019-3-30-.txt
#> ├── 20190331.seq
#> │   ├── 03.0000.XLS
#> │   └── 2019-3-31-.txt
#> └── Data.Rproj

^{Created on 2019-04-02 by the reprex package (v0.2.1.9000)}

Data.zip