Add :sep as an alias for :col_sep

[jsxs0]

This PR adds :sep as a shorter alias for :col_sep to improve usability:

• Adds :sep option that acts as alias for :col_sep
• Reduces keystrokes (3 vs 8) when specifying separator
• Aligns with other data libraries like pandas (pd.read_csv(sep="\t"))
• Maintains backward compatibility
• Raises clear error if both options specified

---

Example usage:

# Old way
CSV.read("data.tsv", col_sep: "\t")

# New way
CSV.read("data.tsv", sep: "\t")

---

The change is motivated by:

1. Making the interface more ergonomic with shorter option name
2. Better alignment with common practices in other data libraries

---

Changes:

1. Adds sep parameter to CSV#initialize
2. Adds error check for both options
3. Uses sep value if provided, otherwise falls back to col_sep

---

Inspired by #272

[kou]

I'm negative this.
If we have a special API for TSV something like #319, we don't need to specify col_sep: explicitly in most use cases?

[kojix2]

I agree with @jsxs0.
I still like the short appearance of sep and do not like the naming of col_sep.
However, I am not going to argue strongly about this. I am not a regular user of the CSV library.

[kou]

OK.
Let's close this.

[kojix2]

Hey, I would like to understand the reason why this PR is closed.

I imagine there are at least 100 people on the planet who would like this change using Ruby's CSV library. Or maybe more. These people may search for pull requests. I feel a kind of obligation to ask for an explanation.

Please understand that I am not asking for the issue to be reopened in order to introduce sep, but rather that I would like to hear your ideas and explanations behind the decision to close it.

Thank you.

---

I don't think that creating a new TSV class is a good enough reason to reject this pull request.
For example, there are many file formats that use “|” as a delimiter.

• Compatibility
• Keeping the code simple
• Consistency with row_sep

There are various possibilities (or the possibility that none of these apply), but the reason is not clear from the outside.

[kou]

Consistency with row_sep

This is the reason.

[kou]

there are many file formats that use “|” as a delimiter.

It will be rare than TSV. Because there is no Pipe-Separated-Values name or something.