Loading from databases in LangChain - Step-by-Step Execution

Concept Flow - Loading from databases

Start

↓

Connect to DB

↓

Run Query

↓

Fetch Results

↓

Create LangChain Docs

↓

Use Docs in Chain

↓

End

This flow shows how LangChain connects to a database, runs a query, fetches results, converts them into documents, and uses them in a chain.

Execution Sample

LangChain

from langchain_community.document_loaders import SQLDatabaseLoader
from langchain_community.utilities import SQLDatabase

db = SQLDatabase.from_uri('sqlite:///example.db')
loader = SQLDatabaseLoader(db=db, query="SELECT * FROM table")
docs = loader.load()
print(len(docs))

This code connects to a SQLite database, loads data from a SQL query as documents, and prints how many documents were loaded.

Execution Table

Step	Action	Evaluation	Result
1	Create SQLDatabase and SQLDatabaseLoader with DB URL	db = SQLDatabase.from_uri('sqlite:///example.db'); loader = SQLDatabaseLoader(db=db, query="SELECT * FROM table")	Loader object ready
2	Call loader.load()	Runs SQL query to fetch rows	Data rows fetched from DB
3	Convert rows to LangChain Document objects	Each row becomes a Document	List of Document objects created
4	Print length of docs	len(docs)	Number of documents printed
5	End	All data loaded and ready for chain	Process complete

💡 All rows fetched and converted to documents, loading finished

Variable Tracker

Variable	Start	After Step 2	After Step 3	Final
loader	None	SQLDatabaseLoader instance	SQLDatabaseLoader instance	SQLDatabaseLoader instance
docs	None	None	List of Document objects	List of Document objects
len(docs)	N/A	N/A	N/A	Number of documents

Key Moments - 2 Insights

Why do we need to convert database rows into Document objects?

LangChain works with Document objects to process text data. The execution_table step 3 shows rows converted to Documents so chains can use them.

What happens if the database URL is incorrect?

The loader creation or load() call will fail to connect or fetch data, stopping execution early before step 2 in the execution_table.

Visual Quiz - 3 Questions

Look at the execution_table, what is the result after step 3?

A An error message

B Raw database rows

C A list of Document objects

D Empty list

💡 Hint

Check the 'Result' column in row for step 3 in execution_table

At which step does the actual database query run?

A Step 1

B Step 2

C Step 3

D Step 4

💡 Hint

Look at the 'Action' and 'Evaluation' columns in execution_table for step 2

If the database has 100 rows, what will len(docs) be after loading?

A 100

B 1

C 0

D Depends on the query

💡 Hint

Variable tracker shows docs is a list of Document objects created from rows

Concept Snapshot

Loading from databases in LangChain:
- Create SQLDatabase.from_uri & SQLDatabaseLoader(query, db)
- Call load() to fetch data
- Rows convert to Document objects
- Use docs in chains for processing
- Handles SQL databases easily

Full Transcript

This visual execution shows how LangChain loads data from a database. First, it creates SQLDatabase from URI and SQLDatabaseLoader with query and db. Then, calling load() runs a SQL query to fetch rows. These rows are converted into Document objects that LangChain can use. Finally, the number of documents is printed. This process lets you use database content in LangChain chains easily.