SAP Hana Driver does not drop prepared statement

[tmilos]

Issue Description

After certain amount of db queries, an error occurs due to exceeded maximum number of prepared statements

Expected Behavior

Single connection should be able to execute infinite number of queries and prepared statements should be closed/dropped when db lib does do it automatically.

Actual Behavior

Error message obtained

Error: exceed maximum number of prepared statements: the number of prepared statements per connection cannot exceed the max statements

Steps to Reproduce

Setup a nestjs with typeorm simple app on SAP hana db, define one entity, and in a loop run inserts for that entity. This brakes after ~34k of inserts

import { InjectRepository } from '@nestjs/typeorm';

  constructor(
    @InjectRepository(Schedule)
    private readonly repo: Repository<Schedule>,
  ) {
  }

for (let i=0; i<=200000; i++) {
      const ent = new Entity({...});
      await this.repo.insert(ent);
}

My Environment

Dependency	Version
Operating System	MacOS BigSur
Node.js version	v12.20.1
Typescript version	v3.8.3
TypeORM version	v0.2.24
@sap/hana-client version	2.6.58
hdb-pool version	0.1.4
@nestjs/typeorm version	7.0.0

Additional Context

SapQueryRunner.query() is doing databaseConnection.prepare() but not doing stmt.drop() in the callback. In my test, after modifying local copy of the file with

// first thing in the prepare callback
statement.drop((err) => {
  if (err) {
    _this.driver.connection.logger.logQueryError(err, query, parameters, _this);
    resolveChain();
    return fail(new index_1.QueryFailedError(query, parameters, err));
  }
})

Note that patch is done on the compiled js in my local node_modules... Typescript would look different ofc.

I am able to go pass the 34k of inserts w/out the error state above. Let it running all the way up to 200k

Please verify error handling stated above is ok - if I'm gonna do the PR

Relevant Database Driver(s)

• aurora-data-api
• aurora-data-api-pg
• better-sqlite3
• cockroachdb
• cordova
• expo
• mongodb
• mysql
• nativescript
• oracle
• postgres
• react-native
• sap
• sqlite
• sqlite-abstract
• sqljs
• sqlserver

Are you willing to resolve this issue by submitting a Pull Request?

• Yes, I have the time, and I know how to start.
• Yes, I have the time, but I don't know how to start. I would need guidance.
• No, I don't have the time, although I believe I could do it if I had the time...
• No, I don't have the time and I wouldn't even know how to start.

[tmilos]

Can we get some update here? If I do the PR, will it be merged?

[1i2hs]

Any updates on this issue? This gives quite a big performance impact on applications using SAP HANA...🥲

[tmilos]

No updates I can see. I could do a PR to fix it, just wanna get some heads up first, not to waist time producing a rejected PR

[imnotjames]

Or write the PR. It's not a huge change to make?

[imnotjames]

I created a PR. Can you test that it corrects the issue?

[tmilos]

Added review, think drop in finally might execute before callback is called. I tested it with drop as first thing in the callback

statement.drop((err) => {
if (err) {
_this.driver.connection.logger.logQueryError(err, query, parameters, _this);
resolveChain();
return fail(new index_1.QueryFailedError(query, parameters, err));
}
})

Sorry, used the phone