Skip to content
datastudyguides

DP-750 · 3.1.4 · Prepare and process data

Choose data format

Choose a data table format, such as Parquet, Delta, CSV, JSON, or Iceberg

Prefer Delta for governed lakehouse tables; use Parquet, CSV, JSON, or Iceberg when interoperability needs justify them.

Separate table format from file format

  • Table format

    Examples
    Delta, Iceberg
    Adds metadata
    Yes
    Transactions
    Table-dependent

  • File format

    Examples
    Parquet, CSV, JSON
    Adds metadata
    No
    Transactions
    No
Format families

A file format describes individual files. A table format adds metadata and table-level behavior across files.

Pick the format by requirement

  • Production Databricks analytical tableDelta
  • Cross-engine Iceberg interoperability is centralIceberg
  • Efficient columnar files are enoughParquet
  • Human-readable interchange mattersCSV or JSON
Format choice

Delta is the Databricks default for governed analytical tables. CSV and JSON are useful at interchange boundaries, not as substitutes for transactional table semantics.

Key points

  • Delta: Default Databricks table format with ACID features
  • Iceberg: Open table format for cross-engine interoperability
  • Parquet: Efficient columnar file format
  • CSV or JSON: Flexible interchange, weaker table semantics

Exam trap

Parquet files alone do not provide Delta transaction log features.

Sources · 2 Microsoft Learn pages Show ▾

Checked on October 4, 2026. Report a problem

Go deeper on Microsoft Learn

Checked against Microsoft Learn on October 4, 2026. Report a problem with this card

How well do you know this?