Hello! :wave: I’m new to GoodData and have a ques...
# gooddata-cloud
a
Hello! 👋 I’m new to GoodData and have a question for the support team (let me know if I should ask in a different channel) - I have two tables in a data source. The data source is AWS Redshift. 1 table was sourced from a PostgreSQL table and the other one was sourced from DynamoDB. However, both are in Redshift as relational SQL tables and being used in the GoodData LDM. Can I connect these two based on shared columns / attributes or is this something that GoodData doesn’t support? Let me know and thanks in advance! 🙏🏼
CC @Faizah Sayyid @Max Donsky
m
Hi Aranta, thank you for reach out us with this interesting question 🙂. If I understood correctly, fortunately, the feature you are asking was released a week ago as Data Blending. It was part of our beta features, but as I mentioned, was completely released to making it usable in production environments. Let me know if this feature is what you are looking for. Cheers, Mauricio.
🙌 1
a
Thanks @Mauricio Cabezas! What exactly do I need to do to enable this? Do I need to upgrade the version or are the latest features pushed automatically for use by all? Also, is it enough to pull both tables into the canvas without any logical connections to any other tables within the same data source? I wrote a SQL query to fetch data from both tables into a custom dataset but data from the DynamoDB table does not show up in the Analyze tab for some reason. Any pointers on how to fix this? As both tables are in Redshift, I expected the query to work as is in GoodData as well and expected to see the same results as what I see in Redshift but in the custom SQL dataset, the DynamoDB table shows 0 results, and the Postgresql table shows all results.
m
Hi Aranta, I will check internally about how and under which condition the feature can be activated and I will back to you with the update. But, take in mind, as a new feature, still has limitations which you can see in the documentation.
🙏 1
a
Sounds good! Thank you.
m
Hi Aranta, just a question for the context - by "sourced from" you mean the data was physically loaded to Redshift from that source? Or is that some kind of link or view? Are both tables in the same schema and accessible to the same user in your Redshift database? If yes, you should be able to use just one data source and connect the tables in your data model without any issues. Just please follow the data modelling recommendations mentioned here: https://university.gooddata.com/tutorials/data-modeling/logical-data-model-introduction/ If your tables are in different databases (or schemas) to connect them you need to use that blending beta feature mentioned by Mauricio above.
a
Hi @Michal Hauzírek, the tables are in Redshift and both have the SQL schema in it. By "sourced from" I mean, data migrated into Redshift from their respective databases (DynamoDB and Postgresql). For GoodData, they should both just be the same type of SQL schema. It seems to break the entire dashboard whenever I add the DynamoDB table. It has about 6000 records and GoodData should be able handle that much right?
i
Hey @Aranta Chatterjee, I don't think that data blending would be required in such case. May I know what type of relationship has you used and what field is it based on? Also, what exactly do you mean by broken dashboards, please? Can you share the error or some screenshot with us, please? Please be aware that update of model might affect existing analytical objects, which are working with attributes/facts/dates affected by such change.
a
Hey @Ivana Gasparekova Can you or Thiago send us a meeting invite at aranta@borderpass.ca and max@borderpass.ca? Max's issues are more urgent (and customer-facing) so we will start by going over that (it'll be quickest to share screen and get answers from a GoodData tech support team) and after that I have some backend system-design related questions for pulling data from multiple data sources (and at a sustainable speed) which will help us make sure our engineering / development work integrates into GoodData seamlessly.
m
Hi @Aranta Chatterjee, First I want to apologize for our delayed response due to the holidays. I have checked, and Thiago is still out of the office until Monday. In the meantime, I would like to investigate this issue further. Could you please revisit the previous questions raised by @Ivana Gasparekova and provide any additional information or screenshot (feel free to use direct messaging for this if it's more convenient), with us? Having this information will ensure that we have clear understanding of the problem (and possible solution), allowing for a more productive discussion on the eventual meeting. Thank you for your cooperation.
a
Hi @Mauricio Cabezas and @Ivana Gasparekova I cannot replicate the broken dashboard issue or provide screenshots for it as that happens in the production workspace which is being used externally. In the LDM, connecting a DynamoDB table which is updated in realtime every few seconds with a Postgresql table leads to all visualizations being blank. So if you can show us how to get these 2 to work together in the meeting, that'll help unblock us. The other issue I can send some screenshots for, in a DM to you.
f
Hi @Aranta Chatterjee, we’ve decided to handle this case via email. I’ve send you an update on your email address with some information to follow-up on this - please check there! 🙂
👍 1
🙏 1