Cockroach: CDC experimantal_avro message encoding

[melskyzy]

Describe the problem
When testing the CDC with experimental_avro, I received the following (while for other formats all the messages are readable):

 NAME 11

, ,NAME 22

J JNAME 37

l lNAME 54

€
€
NAME 64

„
„
NAME 66

¢
¢
NAME 81

¦
¦
NAME 83

Æ
Æ
NAME 99

à
à
NAME 112

   NAME 144

Ê ÊNAME 165

î îNAME 183

ü üNAME 190

Š ŠNAME 197

œ œNAME 206

¼ ¼NAME 222

ê êNAME 245

Ž ŽNAME 263

Ü ÜNAME 302

ð ðNAME 312

ú úNAME 317

š šNAME 333

® ®NAME 343

º ºNAME 349

Î ÎNAME 359

Ô ÔNAME 362

è èNAME 372

ˆ ˆNAME 388

” ”NAME 394

¨ ¨NAME 404

 ÂNAME 417

Ö ÖNAME 427

â âNAME 433

ô ôNAME 442

ö öNAME 443

‚ ‚NAME 449

– –NAME 459

´ ´NAME 474

È ÈNAME 484

¬ ¬NAME 534

° °NAME 536

Ð ÐNAME 552

Ô ÔNAME 554

Ø ØNAME 556

ø øNAME 572

Ž Ž NAME 583

” ” NAME 586

˜ ˜ NAME 588

¶ ¶

To Reproduce

What did you do? Describe in your own words.

• created a changefeed as described in docs.

[knz]

@melskyzy thank you for your inquiry.
Avro data can be encoded using either a binary format or a JSON format. https://avro.apache.org/docs/current/spec.html#Data+Serialization

CockroachDB supports the binary format. This what you see.

Is there anything else we can help you with?

[rolandcrosby]

@melskyzy specifically, this is the binary Avro format encoded according to the conventions of Confluent's schema registry -- the wire format is described here. You can use the kafka-avro-console-consumer utility that comes with the Confluent platform to decode the binary format.

[knz]

@melskyzy I peeked at your github profile and it looks like you are a student at UToronto. Friendly advice from a past teacher, to prepare you to work life: when you feel satisfied that your question has been answered, please write a comment that say "thank you, this answers my question" or something similar. Your peers in the industry will greatly thank you for communicating your status in that way.

For now I am closing this issue. If you want to discuss further, please let us know.