-
Notifications
You must be signed in to change notification settings - Fork 27
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Error when trying to use sqlite3 #35
Comments
Interesting. I think it because the the current directory ( If you write in any other folder say |
If I create a folder manually (eg via right click into sidebar) and try to reference into that I get the same error. Also, saving the db to |
I'm guessing this issue is just carried over from when the kernel was part of the core Jupyterlite repo? jupyterlite/jupyterlite#840 |
@psychemedia leaving here two workarounds:
import sqlite3
con = sqlite3.connect('/tmp/tutorial.db')
cur = con.cursor()
cur.execute("CREATE TABLE movie(title, year, score)")
res = cur.execute("SELECT name FROM sqlite_master")
res.fetchone()
|
@oscarpereztuta Thanks... if |
@psychemedia, SQLite databases opened with the sqlite3 module in Python do persist changes to disk by default. To ensure that changes to the database are persisted to disk, you can set the journal_mode of the database connection to "WAL" (Write-Ahead Logging) mode. Here's an example of how to set the journal_mode to "WAL" in a sqlite3 connection using the SQLite3 Tutorial: import sqlite3
import shutil
# create a connection to a SQLite database file in the /tmp directory
con = sqlite3.connect('/tmp/tutorial.db')
# set the journal_mode to WAL (Write-Ahead Logging) for better performance and reliability
con.execute('PRAGMA journal_mode=WAL;')
# create a cursor object to execute SQL commands
cur = con.cursor()
# create a new table named "movie" with three columns: title, year, and score
cur.execute("CREATE TABLE movie(title, year, score)")
# retrieve the names of all tables in the database
res = cur.execute("SELECT name FROM sqlite_master")
res.fetchone()
# insert data into the "movie" table
cur.execute("""
INSERT INTO movie VALUES
('Monty Python and the Holy Grail', 1975, 8.2),
('And Now for Something Completely Different', 1971, 7.5)
""")
# commit the changes to the database
con.commit()
# close the cursor and connection
res.close()
con.close()
# move the file from /tmp to /drive JupyterLite directory
shutil.move('/tmp/tutorial.db', '/drive/tutorial.db') Then you can download the DB file from JupyterLite and see the result on a web application like this one: https://sqliteviewer.app/ |
@oscarpereztuta Thanks for that. The interaction I am trying to simplify is for a non-expert learner in a distance education setting who just wants to run a notebook in JupyterLite and have changes persisted in local browser storage that they can make use of across multiple sessions. |
If you want to use the home dir, someone would need to figure out what is the missing functionality currently in DriveFS required for Sqlite3 and add it. I don't know enough about this. I see there are different backends for browser storage it could be interesting to try to change the backend and see if the issue persists. Writing the DB in a manually mounted IDBFS (persistent) folder appears to work but that's not very user friendly from a notebook perspective. DriveFS is also able to use IDBFS from Emscripten as a backend, so this does indeed sound like a bug. |
@psychemedia, could you check if this issue has been resolved for you? For me, the issue seems to have gone away: with the versions:
|
I just tried with https://jupyterlite.readthedocs.io/en/stable/_static/lab/index.html and the default Pyodide kernel in a browser with local storage cleared, and now it fails even quicker for me? |
Try our site: https://live.brightway.dev/ ...and alternatively try incognito mode when using the https://jupyterlite.readthedocs.io site. I think it works for me: |
With the jupyterlite demo, chrome incognito, 126.0.6478.127 arm64, Mac M2 14.5: I get the same error on https://live.brightway.dev/lab/index.html |
...yeah, ok - I had remembered that even the import sqlite3
con = sqlite3.connect("tutorial.db")` had thrown an error. I can confirm that res = cur.execute("SELECT name FROM sqlite_master") still does not work for me either. |
@rth, at least I was able to follow your suggestion:
Now that I was able to actually choose different storage drivers (cf. jupyterlite/jupyterlite#1152). I can confirm that the issue persists in both the case of: @jtpio - any idea what the underlying issue might be? We could put some resources towards a fix through UpWork (we would really love to get |
Following the
sqlite3
docs at https://docs.python.org/3/library/sqlite3.html, an error is raised when trying to query a SQLite database:Throws error:
The text was updated successfully, but these errors were encountered: