Read Result Columns by Name

The first available row currently appears as a positional tuple:

(1, 'IT-100', 'Cordless drill', 'tools', 7)

You know that index 1 is the asset tag only because you can see the SELECT list. A caller that writes row[1] quietly depends on that order. This version says what it means:

row["asset_tag"]
row["name"]

SQLite can provide that named access with sqlite3.Row. The setting belongs to the connection and must be in place before the query executes:

connection.row_factory = sqlite3.Row
result = connection.execute("""complete query here""")

Open catalog_db.py. Keep the available-items SQL, result, fetchall(), and the temporary (result, rows) return exactly as they are. Add the row-factory assignment immediately before result = connection.execute(...).

The two objects remain different. result is still a Cursor. Each value inside rows is now a Row. A Row is not a dictionary, although it supports both positions and selected column names. With row = rows[0], all of these refer to values produced by the query:

row["id"]
row["asset_tag"]
row["name"]
row["category"]
row["loan_days"]

Press Run. It makes the same call as before: list_available_items(connection). The outer return remains a tuple, the result type remains Cursor, and the row count remains four. The changed line is concrete:

First row type: Row
First row values: 1 | IT-100 | Cordless drill | tools | 7

The order of those printed values still comes from the exact query, while your Python code can now use their names. That makes a later change to the selected-column order much less likely to give a caller the wrong value silently.

Submit supplies a default connection, so the function itself must configure this step. It also tries an empty available result and checks that you return an empty row list rather than failing at rows[0]. Keep the connection open. You will move this setting to the one place where connections are opened next.