Issue #8154 : Support binary data in bulk load transforms

[mattcasters]

fixes #8154

Hop Binary stream fields (hash keys, raw bytes, BLOBs) can now be bulk loaded on every transform whose engine accepts binary in its bulk protocol. There is no new dialog option: the stream type drives the encoding, the same way PostgreSQL already serializes booleans as t/f.

The previous failure (PGBulkLoader doesn't handle the type Binary) is gone. Text-based loaders call getBinary() and write hex in the format that engine actually reads:

Loader	Encoding
PostgreSQL / Greenplum	\xdeadbeef (bytea hex)
MySQL	lowercase hex + SET col = UNHEX(@col)
Oracle	hex in the data file + HEXTORAW in the SQL*Loader control file (replaces the broken startlob / byte[].toString() path)
Snowflake	hex + BINARY_FORMAT = 'HEX' on COPY
Redshift	hex without \x (VARBYTE)
MonetDB	hex for BLOB
TeraFast	hex for FastLoad BYTE/VARBYTE
CrateDB	hex string (no native binary type; lands in STRING)
SQL Server	unchanged — already sent as byte[]
Vertica	unchanged native bytes; BINARY/VARBINARY now use getBinary()

Out of scope: Doris (it stream-loads a payload field you already built) and the MySQL/SQL Server actions (they load an existing file, not Hop rows).

Tests

Unit tests cover the hex helpers and write paths for PostgreSQL, MySQL, Oracle, Snowflake, Redshift, MonetDB, TeraFast, CrateDB, and Vertica.

Existing integration tests were run successfully for projects database, mssql, vertica, and monetdb. Project cratedb is disabled. None of those suites currently load Binary/BYTEA/VARBINARY columns.

---

Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:

• Run mvn clean install apache-rat:check to make sure basic checks pass. A more thorough check will be performed on your pull request automatically.
• If you have a group of commits related to the same change, please squash your commits into one and force push your branch using git rebase -i.
• Mention the appropriate issue in your description (for example: addresses #123), if applicable.

To make clear that you license your contribution under the Apache License Version 2.0, January 2004
you have to acknowledge this by using the following check-box.

• I hereby declare this contribution to be licensed under the Apache License Version 2.0, January 2004
• In any other case, please file an Apache Individual Contributor License Agreement.

[hansva]

Can we add some IT tests for the databases in the integration tests? Other than that, looks like a solid addition!

[mattcasters]

I couldn't immediately see a non-spammy way of doing it. It's tricky from an IT test-project perspective but I'll try to use /dev/random to generate some sample data in the various IT projects. Another option would perhaps be to use a JavaScript or UDJC transform to generate, perhaps in combination with the new "Data Set Output" transform.

[hansva]

Ah, I see the issue, as we can't include binary data in our source code.
Maybe Base64 encoded text that we converto to binary strings?

[mattcasters]

It's not a bad idea but as you can see we now already have a bunch of prrojects that need the bulk loader test. So we'd be copying files all over the place to make it work properly in the docker containers. Anyway, I don't think generating the binary data is as big of an issue as I thought it was yesterday. I'll figure it out.

[mattcasters]

Added support for binary data in data sets as hex, so a combination of ideas.