Known limitations¶
ChakyBricks 1.0 is a focused, single-user desktop tool. These are the things it does not do, or does with restrictions.
Platform and installation¶
- Windows only. There is no macOS or Linux build yet.
- Unsigned installer. Windows SmartScreen shows "Windows protected your PC"; choose More info → Run anyway.
- Single user, single machine. There are no accounts, sharing, teams or cloud storage connectors. Open the same workspace from one ChakyBricks at a time.
Jobs¶
- Jobs run only while ChakyBricks is running. There is no background service. Turn on Settings → Background → Keep running in the background to keep jobs going after you close the window. Runs that were due while the app was not running are skipped, not made up later, and a run interrupted by closing the app is marked Failed. See Jobs and background mode.
SQL¶
- Limited writes. Only
CREATE [OR REPLACE] TABLE … AS SELECT,INSERT INTO … SELECT|VALUES,DROP TABLEandALTER TABLE … RENAME TOwrite data.UPDATE,DELETE,TRUNCATE,COPY, views, schemas, temporary tables and plainCREATE TABLEare rejected with a message. See SQL editor. - No SQL time travel syntax. Use a table's History tab or
cb.table(name, version=n).
Data¶
- Excel import is limited to
.xlsxfiles of about 50 MB; use CSV or Parquet for anything larger. - Result grids load the first rows of a result (10,000 by default in the SQL editor, 1,000 in notebooks and dashboard tables) together with the exact total; exports contain everything.
Memory and disk¶
- Dashboards load their tiles one query at a time by default, which keeps memory low on very large tables. You can raise this in Settings.
- Query temporary space is capped at the smaller of 50 GB and half of the free space on the workspace's drive. A query that needs more fails with "Not enough disk space for this query".
Python packages¶
- Pinned core packages.
duckdb,deltalake,pyarrow,polars,numpy,pandasandipykernelstay at the versions that ship with ChakyBricks so thatcbkeeps working. Installing a conflicting version is refused with an explanation. - A package installed from Settings → Python packages is importable in an already running notebook only after you restart its kernel. (
%pip installinside a notebook works immediately for packages the kernel has not imported yet.) - Installing packages needs an internet connection.
Found something else? Please open an issue.