I am using a notebook to load an Excel file (downloaded from a website) into a folder in the Files area of a lakehouse - I thought this would be pretty straightforward, but I must be missing something: from datetime import datetime import pandas as pd url = "https://<url_of_excel_file" output_path = "Files/sales_targets/" + datetime.now().strftime("%Y%m%d") # load Excel file from URL and replace spaces in column names df = pd.read_excel(url) df.columns = df.columns.str.replace(' ','') # create directory if it doesn't exist mssparkutils.fs.mkdirs(output_path) df.to_excel(output_path + "/targets.xlsx") Is df.to_excel the correct method here, or should I be using PySpark instead?

Passed DP-600 last Friday - thanks to everyone in this community for your help and support. I'm considering DP-700 next, seeing as there is quite a lot of overlap already with the DP-600 syllabus.

There's a section here for Fabric wins, but could we have one for losses? Or misses? ha! I'm only half kidding. Wanted to share something that I have learned recently about Fabric, specifically related to deployment pipelines. We use deployment pipelines in most of our client projects. Typically, we follow dev, test, prod, but sometimes we use dev, test, qa, prod. We have really enjoyed and benefited from deployment pipelines the past several years before using Fabric. It works wonderfully with semantic models and reports. However, with Fabric data engineering items, not so much. Though, I'm told there's hope that there will be improvements soon. Things I've learned about fabric items when using deployment pipelines 1. dataflows don't work with deployment pipelines. You must manually export the json and import it in your next workspace. This means any changes you make in dev that are ready to move to test have to manually moved over. 2. data pipeline connections have to manually configured whenever you deploy from one workspace to another. There is not currently parameterized connections in pipelines, so you can't setup deployment rules to switch the connection like you can for other items (such as the lakehouse a semantic model points to or a parameterized connection string in a semantic model) 3. SQL tables will migrate from a data warehouse, but the data won't come with them. -- data will need to be manually loaded (or use a notebook or some other automation to pull in the data) 4. Similarly, manually loaded tables in a lakehouse don't get copied over. They will need to be manually created in the new lakehouse (tables created from notebooks can be set to auto create in the new workspace provided that you copy your notebook over as well) 5. shortcuts don't work with deployment pipelines -- they also need to be created manually I'm told that parameterization of data pipeline connections is coming in Q1 2025 and that dataflows are also set to start working in deployment pipelines in Q1 (though they were originally supposed to be available in Q4 2024).

Has anyone experienced problems following the Semantic Link tutorials? I'm getting a 'DatasetNotFoundException' when trying to read the Customer table from the PBIX file (which I renamed to '0_PBIX_CustomerProfitabilitySamplePBIX' prior to upload in my workspace)

Leave a comment with the scariest thing you can say to a Microsoft Fabric Data Engineer / Data Analyst / Data Scientist / Fabric Security Engineer 👀 Scariest saying wins 🎃