Troubleshooting

Stuck? Here are solutions to common problems.

TypeScript Errors

Property 'x' does not exist on type 'Row'

Cause: TypeScript doesn't know the schema of your DataFrame yet.

Solution:

1. Check Inference: Did you load from a CSV? Check if the column exists in df.columns().
2. Cast: If you know it's there:

   df.filter((row: any) => row.x > 10);

3. Better Solution: Use generics (coming soon) or define specific schema.

Argument of type string[] is not assignable...

Cause: You passed an array instead of arguments.

Wrong: df.select(['a', 'b'])Right: df.select('a', 'b')

CSV Parsing

"All my columns are one big string!"

Cause: Wrong delimiter. Solution:

readCsv('file.csv', { delimiter: ';' }); // Try ; or \t

"My numbers are strings!" (e.g., "1,000.00")

Cause: CSV contains formatting chars like commas in numbers. Solution: Clean it after loading:

df.assign('price', row => 
  parseFloat(row.price.replace(',', ''))
);

Runtime Errors

FATAL ERROR: Ineffective mark-compacts near heap limit

Cause: You ran out of RAM trying to load a massive file. Solution:

1. Increase Node/Bun memory: NODE_OPTIONS="--max-old-space-size=4096" bun run script.ts
2. Use scanCsv (LazyFrame) instead of readCsv.

DateTime Issues

"My dates are all invalid"

Cause: The format string doesn't match the input.

Debug:

1. Load as string first (default behavior).
2. Inspect the format: 2023/01/01 is not ISO-8601.
3. Use explicit format:

   readCsv('file.csv', {
       datetime: { columns: { 'date': { format: 'sql' } } }
   })