io.questdb.site.import.readme.txt Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of questdb Show documentation

QuestDB is high performance SQL time series database

The newest version!

This is the default directory for the SQL COPY command and the read_parquet() SQL function.

Drop files here to import data into QuestDB.

You can change the default directory by setting the cairo.sql.copy.root property in server.conf.

See:
    CSV import: https://questdb.io/docs/guides/import-csv/
    Reading external Parquet files: https://questdb.io/docs/reference/function/parquet/
---

This directory also contains a demo parquet file: trades.parquet.
You can use to test the read_parquet() SQL function.
Open the QuestDB web console and run the following SQL query: SELECT * FROM read_parquet('trades.parquet');

You can run aggregation queries on the data in the Parquet file. For example:
SELECT MIN(price), MAX(price)
FROM read_parquet('trades.parquet');

Data in the demo Parquet file are sorted by timestamp. This allows for efficient time-series queries,
but you need to specify the timestamp column when reading the Parquet file. For example, to sample the data by 1 day:
WITH trades AS (
    (SELECT * FROM read_parquet('trades.parquet')) TIMESTAMP(timestamp)
)
SELECT timestamp, MIN(price), MAX(price), FIRST(price) AS open, LAST(price) AS close FROM trades
SAMPLE BY 1d;

And last but not least, you can create a regular table and import the data from the Parquet file into it:
CREATE TABLE trades AS
(SELECT symbol::symbol AS symbol, side::symbol AS side, price, amount, timestamp FROM read_parquet('trades.parquet'))
TIMESTAMP(timestamp)
PARTITION BY MONTH WAL;