# Use Rasa Conversational Data for Retraining

**URL:** <https://forum.rasa.com/t/use-rasa-conversational-data-for-retraining/56921>\
**Category:** Rasa Open Source\
**Created:** [January 20, 2023, 10:08am UTC](https://forum.rasa.com/t/use-rasa-conversational-data-for-retraining/56921 "2023-01-20T10:08:39Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![rish5](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/rish5/32/22301_2.png) [@rish5](https://forum.rasa.com/u/rish5)\
**Post date:** [January 20, 2023, 10:08am UTC](https://forum.rasa.com/t/use-rasa-conversational-data-for-retraining/56921/1 "2023-01-20T10:08:39Z")

</div>

I saved the conversational data using a tracker\_store from endpoints.yml file. It looks like this: - tracker\_store: type: “sql” dialect: “postgresql” url: “postgresql://user:password@URL:PORT/database” port: 5432 db: database password: password Now I want to use the data from postgresql for training. Is it possible because manual implementation of such a large conversation is not feasible? I have used csv data but it says unknown data format for csv file. For conversion I have used rasa data convert on rasa 3.4.0 version but got unknown data format error. What pre-processing needs to be done on for rasa training?