Import catalog (SaaS mode)

Beyond connecting external catalogs directly to Data Explorer (Refer: External catalog), you can import catalog through a CSV file as described in this article.

This procedure consists of creating a table with schema details and importing the CSV file. The table creation is needed only for the first import. The following CSV file is used to describe the process.

file_path,gt_class,gt_box,pd_class,pd_box,pd_score
VOCtrainval_11-May-2012/VOCdevkit/VOC2012/JPEGImages/2011_000753.jpg,car,"{'x_left': 0.812666902937676, 'x_right': 0.9925274478643895, 'y_top': 0.5254668930645845, 'y_bottom': 0.9725518810661693}","","",""

Create table

1. Navigate to 'Data -> Repo -> DataSets' and click on the 3-dots icon on the dataset card corresponding to the dataset within which the table needs to be created.
   
   Dataset Action button
   
2. Enter the table name and table description.
3. Enter the column names and the data type for each column. The column names must match the column names in the first(header) line of the CSV file. The Column type should be a valid SQL type. Click 'Submit' after filling in all the column details.
   

Import catalog

1. Navigate to the dataset catalog page. The newly created table will be listed on this page. Click on the 3 dots and select 'Import Files'.
   
   
2. Select one or more CSV files and click 'Upload'.
   
   
3. Once upload is complete, click 'Submit'.
   
4. This creates an import job that runs asynchronously. The status of the import job is available on the 'Import Jobs' page, accessible from the 3 dots icon against the table.
   
5. The Import Jobs page opens up with a listing of all import jobs and their status.
   
6. Clicking on the entry in the JOB ID column provides detailed status about the Import job. Along with the status, this page has details about the number of rows ingested, and the number of rows skipped (if any)