bugfix: format schema cache is not cleared after the schema is dropped

[zliang-min]

PR checklist:

• Did you run ClangFormat ?
• Did you separate headers to a different section in existing community code base ?
• Did you surround proton: starts/ends for new code in existing community code base ?

Please write user-readable short description of the changes:

In this PR, I have ported a few changes related to ProtobufSchemas from CH repo for cleaning up the Protobuf schema cache (please check the Commits tag for what are ported). And based on those changes, I added the logic to DROP FORMAT SCHEMA and CREATE OR REPLACE FORMAT SCHEMA to clear theProtobuf schema cache.

Here is a use case that is blocked by this bug:

1. Create a protobuf schema, for example

   CREATE FORMAT SCHEMA foo as '
   syntax = "proto3";
   
   message Foo {
     string a = 1;
     int64 b = 2;
   }
   ' TYPE Protobuf;

2. Use the schema with a Kafka external stream:

   CREATE EXTERNAL STREAM foo (
   a string,
   b int64
   ) SETTINGS type='kafka',...,data_format='ProtobufSingle',format_schema='foo:Foo';
   
   SELECT * FROM foo;

   This works fine.
3. Replace the schema with a new one ( like the schema is evolved ):

   CREATE FORMAT SCHEMA foo as '
   syntax = "proto3";
   
   message Foo {
     string a = 1;
     int64 b = 2;
     string c = 3;
   }
   ' TYPE Protobuf;

   In this example, we added a new field c.
4. Use the new schema with an evolved external stream

   CREATE EXTERNAL STREAM foo (
   a string,
   b int64,
   c string
   ) SETTINGS type='kafka',...,data_format='ProtobufSingle',format_schema='foo:Foo';
   
   SELECT * FROM foo;

   This time, the values of column c will always be empty, because the schema is cached, and even after we re-create it, it still uses the old cached schema.

[chenziliang]

LGTM in general

[jovezhong]

(Jove Github Bot) added it to the current sprint.

[jovezhong]

(Jove Github Bot) moved this ticket out of the GitHub project(up to 1200 tickets for one project).