ARROW-18421: [C++][ORC] Add accessor for stripe information in reader

[LouisClt]

See https://issues.apache.org/jira/browse/ARROW-18421

[github-actions]

https://issues.apache.org/jira/browse/ARROW-18421

[github-actions]

⚠️ Ticket has not been started in JIRA, please click 'Start Progress'.

[pitrou]

Why not expose the full stripe information? We don't want to create tons of little accessors for each different piece of information.

[LouisClt]

Yes, why not.
The structure storing the different informations available is located only in the implementation :

struct StripeInformation {
  uint64_t offset;
  uint64_t length;
  uint64_t num_rows;
  uint64_t first_row_of_stripe;
};

The offset is the number of bytes since the beginning of the file, in bytes, the length is the size of the stripe in bytes, num_rows is the number of rows in the stripe, and "first_row_of_strip" is the index of the first row in the stripe (thus the sum of the num_rows of all the preceding stripes).
We could just put the definition of this struct in the header, with comments on the use of each of these members, and return this struct in the getter.
But maybe some members would benefit from a renaming...

[pitrou]

We can change it slightly and expose it as:

struct StripeInformation {
  int64_t offset;
  int64_t length;
  int64_t num_rows;
  int64_t first_row_id;
};

(note the use of signed integers to match our API conventions better)

[LouisClt]

Ok, that seems fine to me. I will make the changes then.
A prototype like this would be OK ?

StripeInformation GetStripeInformation(int64_t stripe);

[pitrou]

Yes, that sounds fine.

[LouisClt]

So this seems to be good now. I added the "StripeInformation" class in the header, and modified the implementation accordingly.
I added one more small test on that.

[pitrou]

+1, thank you @LouisClt