The Duck Archives¶

Long-term-memory

Before you launch, you need to create a backup of all important data. You head for a duck with a fantastic memory and tell them everything.

The duck speaks a strange language though - SQL.

Installing DuckDB¶

DuckDB is a Python library. Install it with:

!pip install duckdb

Read Data¶

Like pandas, duckdb can read data directly from CSV files and other formats.

import duckdb

panda_sector = duckdb.read_csv("panda_sector.csv", index_col=0)
df = panda_sector.to_df()

SQL Queries¶

With SQL, you can answer questions to the data:

duckdb.sql("SELECT * FROM 'panda_sector.csv' WHERE class='M'")
duckdb.sql("SELECT name,size FROM 'panda_sector.csv' WHERE class='M' AND size> 10 ORDER BY name LIMIT 10")

Note that you can refer to tables by the file path or by a Python variable name:

duckdb.sql("SELECT * FROM panda_sector LIMIT 5").df()

Store the Database in a file¶

So far, eveything was stored in memory only. Lets' make it persistent:

con = duckdb.connect("starbase.db")
con.read_csv("panda_sector.csv")
con.sql("CREATE TABLE dummy AS SELECT * FROM 'panda_sector.csv' LIMIT 5")
con.close()

Now you should see a file starbase.db .

con = duckdb.connect("file.db")
con.sql("SELECT * FROM dummy")