#339. Use more effective record count when JDBC source is used

[VladimirRybalko]

No description provided.

[github-actions]

Unit Test Coverage

File	Coverage [84.34%]	🍏
TableReaderJdbc.scala	84.34%	🍏

Total Project Coverage	82.23%	🍏

[yruslan]

Thanks for the PR! Looks good.

[yruslan]

Alternatively, we don't really need Spark to execute a count() query.

You can also do it like thins:

  private[core] def getCountForSql(sql: String, infoDateBegin: LocalDate, infoDateEnd: LocalDate): Long = {
    val filteredSql = TableReaderJdbcNative.getFilteredSql(sql, infoDateBegin, infoDateEnd)
    val countSql = s"SELECT COUNT(*) FROM ($filteredSql)"
    var count = 0L

    JdbcNativeUtils.withResultSet(jdbcUrlSelector, countSql, jdbcRetries) { rs =>
      if (!rs.next())
        throw new IllegalStateException(s"No rows returned by the count query: $countSql")
      else {
        count = rs.getLong(1)
      }
    }

    count

But this requires rebasing against the latest main

[yruslan]

Looks good!