JDBC input does not work with PostgresSql numeric column.

[marekduciuc]

• Version: 4.3.19 with logstash 7.5.0

• Operating System: ALL

• Config File (if you have sensitive info, please remove it): important art is:
  input
  {
  jdbc
  {
  jdbc_driver_class => "org.postgresql.Driver"
  jdbc_connection_string => "jdbc:postgresql://host:5432/schema"
  jdbc_user => "user"
  jdbc_password => "pass"
  statement => "select
  id as id,
  log_record as log_record,
  creation_time as creation_time,
  to_char(creation_time, 'YYYY.MM') as log_date
  from sz_audit_log
  where
  id > :sql_last_value AND
  id < ((select MAX(id) from sz_audit_log where id > :sql_last_value) - 100)
  order by id
  "
  use_column_value => true
  tracking_column => "id"
  tracking_column_type => "numeric"
  jdbc_paging_enabled => true
  jdbc_page_size => "1000000"
  schedule => "3,18,33,48 * * * *"
  last_run_metadata_path => "/srv/logstash-data/.sz_audit_log_jdbc_last_run"
  }
  }

• Sample Data: NA

• Steps to Reproduce:
  User has configuration as above:

Tracking column is defined in Postgres as "id | numeric(19,0) | "
When they run pipeline last value is stored in last_run_metadata_path as following:

--- !ruby/object:BigDecimal '0:0.4136727923e10'

This caused an issue when the pipeline is run again as logstash refuse o use this value and start grabbing values from 1st again.
There is not any error in logs it s just hat logstash does not use the value stored in last_run_metadata_path

This works correctly with plugin 4.3.13