andrus.adamchik.org
Java open-source & ObjectStyle
Making Java a viable option for data processing / analysis at DFLib.org
Apache Software Foundation member / ex VP
168 posts
303 followers
183 following
Prolific Poster
Conversation Starter
comment in response to
post
Try googling (or whatever the term is for ChatGPT π) for "Jupyter keyboard shortcuts". Notebook keyboard navigation is awesome and makes you very productive (inserting cells above and below, changing cell type between code and markdown, etc.)
comment in response to
post
The "TAB" key should do it.
comment in response to
post
Ok, the first version of the docs just got published at the link above.
comment in response to
post
Thanks! Yeah, we'll prioritize filling this gap in the docs. For now, look up the class called "Excel" and its static methods. E.g.:
// load all sheets
Map<String, DataFrame> data = Excel.load("my.xlsx");
// load a single sheet
DataFrame df = Excel.loadSheet("my.xlsx", "Sheet 1");
comment in response to
post
Very nice! I resisted the urge to make β$β a method name in DFLib for expression shortcuts, but here it looks very natural π
comment in response to
post
Thatβs assuming human developers catch up with all the new stuff in each Java version π
comment in response to
post
DevNexus is off to a great start! π
comment in response to
post
Those are like Docker containers - you need a dedicated one for every process
comment in response to
post
Great, going to try it. Learned something new today π
comment in response to
post
Yeah smth like this. Not fully sure how Binder works, but I assume it would require Java kernel install somewhere behind the scenes
comment in response to
post
Being one of the Java kernel developers, Iβd like to see better exposure of kernels for different language (well, Java in partcular π) . Eg here jupyter.org/try#kernels and this wonβt even require any noticeable dollar spending
comment in response to
post
I think all platforms turning into store fronts (a nicer term than "enshitification" but same phenomenon really) is a likelier reason than simply fragmentation.
comment in response to
post
Blockchain is just blockchain. But LLMs are so complex that nobody could say for sure they couldn't become AGI. So everyone fully committed out of FOMO. Nobody figure out AGI, corps got lefts with ML models they can't find any real use for, and investors really don't want to be left bagholding. 2/2
comment in response to
post
Pretty much, but if you are working in Java that has no culture of either REPL or notebooks, it is also a gateway drug to that style of programming
comment in response to
post
Are you still a programmer if you start counting indices from "1"? π
comment in response to
post
One reason I noticed why tech docs are not searchable is they are often long books published as a single HTML page. So Google doesnβt rank them well for specific topics. I am considering breaking mine into many pages to test the hypothesis
comment in response to
post
Yes, pretty much. Fundamentally, it is those 3 things:
1. API to control rendering of certain Java objects in a certain way that a notebook code or dependency can tap into
2. Fixed-width character rendering (DataFrame data and such)
3. HTML/CSS/JS rendering (charts)
comment in response to
post
Red Hat was acquired by IBM in 2019, so I guess it is just shuffling things around internally?
comment in response to
post
Didn't know about this lib. Gatherers are nice, but writing own gatherers is unpleasant (just like own collectors). Gatherers4J seems to fill this gap quite nicely. Also looks like the link is broken. The right one is github.com/tginsberg/ga...