jupyterlab-sql-editor
jupyterlab-sql-editor copied to clipboard
CybercentreCanada
Not clear how to run this on windows for jupyterlab 3 or 4
I am not clear how to run this inside jupyterlab 3 or 4 and actually get %%sparksql magic to work? I have both jupyterlab 4.2.4 and 3.6.7 running, without errors in the output or the tmp log file, but it says cell magic not found when i try %%sparksql magic
Current version works with JupyterLab 4. I just released slightly improved logging for loading errors. In general you need to have sql-language-server installed with npm -g (node 20 recommended), node available in your notebook kernel's path and for Python you need jupyterlab-lsp>=5.0.0 jupyter-lsp>=2.2.0 python-lsp-server[all].
I admittedly exclusively work on Linux but I'll give it a try on Windows and see how it goes. Your Jupyter logs have a good clue. I'll get back to you soon @Paradox137
I made some changes in version 1.4.5 which should make it possible to use on Windows now. Let me know if it works for you or if you run into other issues @Paradox137.
@cccs-nik hi, servers works, but... Py4JJavaError always on calls and syntax don`t highlight
Tag me if you need more log information, because I really don't even know what the problem could be, having tried many solutions on the net - nothing helps
Maybe i need again dawngrade python version but i don`t see: "RuntimeError: Python in worker has different version..." in logs
@Paradox137 Hey sorry was a little busy last week. I can't really tell from your exception what the issue is. The error message looks a bit generic, not sure if you have better logs somewhere else or not. First I would make sure you have compatible Java and Python versions for PySpark 3.5.5. That would be Python 3.8 to 3.11 and Java 8/11/17 if I'm not mistaken. You probably need multiLine=True to your json read given your schema says _corrupt_record: df = spark.read.json("sample1.json", multiLine=True). Syntax highlighting should be working given the LSP servers window you screenshotted.