read_table, header flag. false interpreted as 0.

[cyrusmaher]

For read_table, Would it be possible to make "header=None" and "header=False" have the same behavior? Interpreting "False" as saying that the first line is the header has tricked me a couple times...

You might have a good point but we can't break that much code without a stronger reason.

Sorry, that's what maintaining a library that has lots of users and existing code is like.

I spoke too soon. The docstring leaves that behavior undefined and what you suggest makes sense.
We'll talk this over for 0.14.0. I think we can do this.

A the very least, raising on bool with a reasonable error message will
eliminate the hazard and not cause much pain for existing code that relies
on undocumented behavior.

[cyrusmaher]

Perfect. If you guys could do it sooner, all the better. In R's read.table, header=False gives pandas' header=None behavior. With the pitfalls of language slurring added to the fine False/None distinction, my feeling is raising an error will cut down on unintended behavior with comparatively little inconvenience...

That's been there for a while. I don't think there's any time pressure.
our next release is a bugfix release and this could conceivably break
someone's ill-written code. It'll wait till 0.14.0. You're welcome to use master
before 0.14.0 is actually released. It's usually perfectly usable.

If we don't sort this within a month or so, ping us.

[cyrusmaher]

Sounds good. Thanks for taking a look!

On Sun, Jan 26, 2014 at 9:05 PM, y-p notifications@github.com wrote:

That's been there for a while. I don't think there's any time pressure.
our next release is a bugfix release and this could conceivably break
someone ill-written code. It'll wait till 0.14.0. You're welcome to use
master
before 0.14.0 is actually released. It's usually perfectly usable.

If we don't sort this within a month or so, ping us.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/6113#issuecomment-33342316
.