Skip to content

Importing data

ChakyBricks stores every table as a Delta table in your workspace. Importing converts a file into such a table; the original file is never changed or needed afterwards.

Supported files

Format Extensions
CSV .csv, .txt
Tab-separated .tsv, .tab
Parquet .parquet, .pq
JSON and newline-delimited JSON .json, .ndjson, .jsonl
Excel .xlsx (files up to 50 MB)

CSV, TSV and Parquet files of several gigabytes are fine: the import streams the data through DuckDB into Delta in the background, so it does not need to fit in memory and does not block the app.

The import wizard

Start it from Home → Import data, from the Catalog, or from the command palette (Ctrl+K). It has three steps: pick a file, configure the import, and watch the progress.

Configure lets you set:

  • Schema and Table name — where the table goes. You can create a new schema on the spot. Names are lowercase letters, digits and underscores.
  • Table already exists — if the name is taken, choose Append rows or Replace table.
  • Partition by — optional; pick a column with few distinct values, such as a date or a region, to split the table into folders. Partitioning is not available when appending.
  • Delimiter and First row is a header — for CSV and TSV files (the delimiter can be auto-detected, or set to comma, semicolon, tab or pipe).
  • Columns — every column can be included or left out, renamed and given another type. The types offered are VARCHAR, BOOLEAN, INTEGER, BIGINT, DOUBLE, DECIMAL(18,2), DATE, TIMESTAMP and TIMESTAMP WITH TIME ZONE.
  • Preview — the first rows are shown with your settings applied, and update as you change them.

Column names

Column names are cleaned so they work unquoted in both SQL and Python: letters of any script (Arabic, accented Latin and others), digits and underscores are kept, everything else becomes an underscore, names are lowercased, and duplicates get a suffix. A name that starts with a digit, or that is a reserved word, gets a c_ prefix. You can still edit every name in the wizard.

Progress and cancel

The progress step shows the rows written so far. You can close the dialog and keep working: a notification appears when the import finishes. Cancel stops the import; a table that was being created is removed again. Only one import runs at a time; others wait in the queue.

The catalog

The Catalog page lists schemas, tables and columns with type icons. Click a table to open its page:

  • Schema — column names and types.
  • Sample data — the first rows.
  • History — the Delta version history (what was written, when). Pick an older version to see a sample of the table as it was then (time travel).
  • Details — full name, location in the workspace, where it was imported from, creation and update times, and the format (Delta Lake).

From the table page you can Open in SQL, Open in notebook, rename the table, or drop it. Dropping deletes the data.

Time travel

Every write creates a new Delta version. In a notebook, read an older version with cb.table("sales", version=3); in the app, use the History tab. See the cb reference.

Creating tables another way

You can also create tables from queries: run CREATE TABLE … AS SELECT in the SQL editor, use Save as table on a result, or call cb.save(df, "name") in a notebook.