png.png
# gooddata-platforms
Sherry
03/03/2022, 4:32 AMpng.png
t
Tomas Muchka
03/03/2022, 7:52 AMHello @Sherry, our hosted version currently does not support file-oriented Data Sources such as S3 or Azure Blob for modeling. The recommended way how to connect your CSV/parquet files stored in S3 is to:
1. Download the files locally.
2. Drop the downloaded files one by one into the modeller window. This will create the model.
Tomas Muchka
03/03/2022, 7:54 AMIf you would like to automate data refresh from the S3, then please go into the Load section (top left part of your screenshot) and set a recurrent data load from your S3 Data Source.
s
Sherry
03/03/2022, 4:23 PMHi @Tomas Muchka I had downloaded the files locally and dropped it into the modeler window, and would like to automate data refresh for new data, but even after setting a scheduled data load from S3, I do not see the analysis refresh with new data. I even see the data load was successful.
t
Tomas Muchka
03/03/2022, 6:17 PMHello @Sherry, thank you for the information. Does the “last loaded” column update as you run the scheduled load? It might be that the load does not take your locally created datasets into account for some reason.
Tomas Muchka
03/03/2022, 6:18 PMPlease check if the schedule says “All datasets” and if the Last loaded column updates according to the expectations
s
Sherry
03/03/2022, 6:25 PMIt looks like as you described here
Sherry
03/03/2022, 6:26 PMThe analysis is still based on the csv file that I uploaded under the datasets load section, and none of the new data from the scheduled loads aws is being updated on the analysis
t
Tomas Muchka
03/03/2022, 6:31 PMGoodData provides a run log to all scheduled loads. You can check there if everything is mapped correctly. Open the “…” menu next to “Run now” and select view run log option. There is a link to View log then.
Tomas Muchka
03/03/2022, 6:35 PMAdditionally you can check if the dataset’s load configuration (model tab > dataset details) is according to the expectations. If it is mapped properly and if all advanced load features such as “Incremental loads” are turned off.
s
Sherry
03/03/2022, 7:16 PMyeah... it seems like everything is as you described...
t
Tomas Muchka
03/03/2022, 7:35 PMEven the load log shows the dataset fields are properly mapped, yet the analytics shows old data?
s
Sherry
03/03/2022, 8:41 PMyes
Sherry
03/03/2022, 8:46 PMIs the Source Column suppose to match the column on the aws s3 or the csv dataset that I had uploaded?
t
Tomas Muchka
03/04/2022, 8:38 AMIt should match the name of the column in your S3 bucket. These should however be the same if you have downloaded the CSV files from S3 and the uploaded them locally into the model.
s
Sherry
03/06/2022, 7:18 PMI had re-uploaded the CSV file to the model and so the Name and Source Column are now matching, however, the analysis is still not updated with the most recent data from 2022.
m
Moises Morales
03/07/2022, 3:56 PMHi Sherry,
I noticed you also created a Community Post regarding this issue. Could you please answer my questions there? TIA.