Write can be called multiple times, so adding a way to prevent writing headers

[jvelasques]

For example, you can be chunk reading documents from an Elastic Search cluster, and writing as you process them.

[joaomatossilva]

Hello..

First of all. Thank you very much for the contribution. It's an awesome feeling to know that my project is actually useful for someone.

So.. no worries about the build. That's something that I need to fix.

About the change itself.
Fist, I would love to have some tests to cover your new feature ;)

Second, I'm not so sure that we need to change the public signature of the Write method in order to be called multiple times.
What if we keep that writeHeaders flag in a private field? that way we would force to just write the headers once, subsequent calls if the writeHeaders would be false, it would skip it.

What do you think?

[jvelasques]

Hi there!

I was actually reluctant to do that just because it changes the flow. Not sure who'd want to write headers in the middle of the file though, so the change does make sense. I'll change this PR!

Thanks!

[joaomatossilva]

Not sure who'd want to write headers in the middle of the file

Yup. My point exactly.

[jvelasques]

@joaomatossilva changed! Hopefully the test makes sense 😅

Thanks!

[diogogcunha]

Just as a nice extra info on this conversation, we were able to use this to stream an Elastic Search paged result into a CSV file directly on Azure Blob never storing anything on the local file system.
It worked really nicely! Congrats on this project

[joaomatossilva]

I've fixed (or so I think) the version tag on master. If you could rebase your change on top of it would be great.

Call me on Whatzapp if you need help doing it ;)