Support converting zero bytes and high-bit-set bytes to octal sequences for VARBINARY

[Obsecurus]

PostgreSQL has a function encode(data, 'escape') where data is a bytea.
https://www.postgresql.org/docs/9.4/functions-binarystring.html

Currently when using sqlline w/ the Athena JDBC the default output for VARBINARY is to convert as hex. It would be extremely useful the have the ability to support this output mode directly in sqlline for any VARBINARY column when set.

[snuyanzin]

@Obsecurus thank you for the raised issue
Not sure that this covers all your needs but anyway there is escapeOutput property which also allows to do escaping.
!set escapeOutput true

[Obsecurus]

we've tried escapeOutput but it does not generate the desired results. We've use the Athena JAR with raw Python and get the desired result. I imagine anyone working with binary data containing ascii would want this mode. I suspect sqlline doesn't have this rendering mode with any of the drivers.

[Obsecurus]

Here is an example select on a VARBINARY column from Athena when using the Python pyathenajdbc library which is a wrapper for the Athena JAR. Notice how the ASCII converts but other other bytes are left as-is:

b'\x03\x00\x00+&\xe0\x00\x00\x00\x00\x00Cookie: mstshash=hello\r\n\x01\x00\x08\x00\x03\x00\x00\x00'

[snuyanzin]

yes, I understand what you mean.
currently i see 2 ways:

1. Use existing java byte array to string + as it is just a byte representation inside db it makes sense to specify encoding e.g. via property (as there could be ascii, utf-8, utf-16 and etc.). However all non-mapped symbols will be just replaced by default with \uFFFD (java's default while conversion from byte array)
2. To do replacement with something depending on byte value need extra efforts. + Also makes sense to have a property to define format e.g. octal sequence, hex sequence and so on