Skip to main content

Accumulate CSVs Command

Roll a new month's merged transactions into the cumulative yearly dataset.

Arguments​

  • $1: Month directory path

Input​

  • New month's file: <month-dir>/agentic_merged_transactions.csv
  • Existing cumulative (if exists): <parent>/agentic_cumulative_dataset_<year>.csv

Output​

  • Updated cumulative: <parent>/agentic_cumulative_dataset_<year>.csv

Accumulation Rules​

First Month​

If cumulative doesn't exist:

  1. Copy new month's merged transactions
  2. Create cumulative file

Subsequent Months​

  1. Read existing cumulative
  2. Read new month's data
  3. Verify no date overlap
  4. Concatenate new data
  5. Sort by date descending
  6. Write updated cumulative

Validation​

  • No date overlap between months
  • Warn if gap between months
  • Check for exact duplicates

Workflow​

  1. Parse directory path from $1
  2. Read new month's agentic_merged_transactions.csv
  3. Check if cumulative exists in parent
  4. Merge with existing data
  5. Sort by date descending
  6. Write cumulative file
  7. Stop hook validates output

Usage​

/finance-accumulate data/mock_dataset_feb_1st_2026/