mercury_csv is a Mercury library for reading
CSV data from character streams.
- follows RFC 4180 where possible
- can optionally trim whitespace from fields
- handles CRLF line endings
- allows blank fields
- supports reading fields either as strings (raw) or as Mercury standard library types (typed).
- the "typed" interface allows an arbitrary number of actions (e.g. transformations, validity checks etc) to be performed on each field after it is read.
- allows limits to be imposed on number of characters in a field and the number of fields in a record
- supports the presence of an optional header record
- field delimiter character can be selected by the user
- can optionally ignore blank lines
- can optionally ignore trailing fields
mercury_csv is licensed under a simple 2-clause BSD style license.
See the file COPYING for details.
Check the values in the file Make.options to see if they agree with your system, then do:
$ make install
You can also override values in Make.options on the command line, for example
$ make INSTALL_PREFIX=/foo/bar install
causes the library to be installed in the directory /foo/bar.
To run the regression test suite, do:
$ make runtests
- write some samples
- write some more documentation
- support CSV writers as well as readers
The code on the master branch is not compatible with Mercury 22.01.X.
If you require a version of mercury_csv that works with Mercury 22.01.X,
then checkout the mercury_22_01 branch.
The code on the master branch is not compatible with Mercury 20.06.X.
If you require a version of mercury_csv that works with Mercury 20.06.X,
then checkout the mercury_20_06 branch.
The code on the master branch is not compatible with Mercury 20.01.X.
If you require a version of mercury_csv that works with Mercury 20.01.X,
then checkout the mercury_20_01 branch.
Julien Fischer juliensf@gmail.com