Dataprep数组类型导出CSV异常：UI显示正常但输出值无效

Let me break down why this happens and how to fix it—this is a super common gotcha with CSV exports and nested data types like arrays.

Core Reason: CSV Doesn't Natively Support Arrays #

First off, remember that CSV is a plain-text, flat file format—it has no built-in way to represent arrays or nested structures. The Dataprep UI is designed to visualize complex data types (like arrays) in a human-friendly way, but when you export to CSV, Dataprep has to convert that array into a single text value. If you don't explicitly tell it how to do that, it might default to a raw, unreadable representation (like [Ljava.lang.String;@123abc—that's just Java's internal object reference, not actual array content).

Common Fixes to Align CSV Output with UI Display #

Here are the most straightforward solutions to make your CSV output match what you see in the Dataprep UI:

1. Convert Arrays to Delimited Strings
   Use Dataprep's join function to turn your array into a single string with a separator of your choice (pick one that won't conflict with your data, like ; or |).

   Example derived column formula:

   join(';', your_array_column_name)
   

   This will turn an array like ["apple", "banana", "cherry"] into apple;banana;cherry, which exports cleanly to CSV as a single cell value.

2. Serialize Arrays to JSON
   If you want to preserve the array structure for later parsing, use the toJson function to convert the array into a standard JSON string.

   Example derived column formula:

   toJson(your_array_column_name)
   

   This will output ["apple", "banana", "cherry"] as a single quoted string in CSV, which is easy to parse back into an array later.

3. Check Export Configuration for Flattening
   Double-check your Dataprep export settings—if you accidentally enabled the "Flatten arrays" option, it will split each array element into its own row, which can make your original array column look like it has missing/invalid values. Disable this option if you want to keep the array intact in a single cell.

Quick Example Workflow #

1. In your Dataprep recipe, add a new derived column using either join or toJson on your array column.
2. Remove the original array column from your output (or keep it if you need it for other steps).
3. Run the export to CSV again—your array data should now match what you saw in the UI.

内容的提问来源于stack exchange，提问作者Warisa