ClintWeathers

The Main Idea

Get a datascience environment built on a VM.

In theory, this gives me a few things:

• The umph of my overbuilt desktop machine minus a few gig of ram and a couple cores
• The ability to SSH/VNC/RDP in and just use my chromebook as a cheap thin client when Im out and about.
• Experience dealing with the install and usage of Drill, Zeppelin, and some other tools.

So far I've:

1. Downloaded the Bionic ubuntu iso

ClintWeathers

# List unique values in a DataFrame column
# h/t @makmanalp for the updated syntax!
df['Column Name'].unique()

# Convert Series datatype to numeric (will error if column has non-numeric values)
# h/t @makmanalp
pd.to_numeric(df['Column Name'])

# Convert Series datatype to numeric, changing non-numeric values to NaN
# h/t @makmanalp for the updated syntax!

ClintWeathers

Most of this is just following the steps here:

http://sherifsoliman.com/2016/01/11/theming-ipython-jupyter-notebook/

On a recent Juypter version, open up the custom.css in your favorite editor which should be Sublime Text 3 or Vim and definitely not emacs.

~/.jupyter/custom/custom.css (on my VM this lives in C:\Users\username\.ipython\profile_default\static\custom -- YMMV)

Do a search for kernel_busy_icon

ClintWeathers

library(dplyr); library(magrittr)

#To Create The Dataframe of Fake Orders

set.seed(867.5309)

onums <- paste("AA", as.integer(runif(20) * 1e8), sep = "")
prices <- round((runif(20) * 1000), 2)
orders <- as.data.frame(cbind(onums, prices))

ClintWeathers

> may_2015 <-
+   read_csv(
+     "Z:/data_CSVs/thingsandstuff_files/may2015.csv"
+   )
Assertion failed!

Program: C:\Users\ahm0\Documents\RStudio\bin\rsession.exe
File: d:/RCompile/CRANpkg/lib/3.2/BH/include/boost/spirit/home/support/detail/pow10.hpp, Line 87

Expression: dim < sizeof(exponents)/sizeof(double)

ClintWeathers

# Started from here, an example from Hadley Wickham in the google groups https://groups.google.com/forum/#!topic/ggvis/AJZCdjFcNaE. 
# Using R 3.2.1 x86_64, ggvis 0.4.2. 

cocaine %>%
  ggvis( ~ potency, ~ price) %>%
  filter(state %in% eval(input_checkboxgroup(c("IA", "MN", "WI"),
                                             selected = "WI"))) %>%
  layer_points(fill = ~ state, size := 50, opacity := 0.5) 
  
# This works fine and evals the way one would expect.

ClintWeathers

### Trying to create new dataframes from dplyr 0.4.3 functions using R 3.2.2.


#Very Simple Example:
paste("cars_with", 6, "cylinders", sep = "_") <- filter(mtcars, cyl == 6)

#What I want there is a dataframe named "cars_with_6_cylinders" that has the results of this:
filter(mtcars, cyl == 6)

#But instead, what I get is this:

ClintWeathers

---
title: "Testing1"
author: "JCW"
date: "November 19, 2015"
output: 
  html_document: 
    highlight: haddock
    theme: cosmo

---

ClintWeathers

library(RODBC)
library(RODBCext)
dbconn <- odbcConnect(dsn = "dbname", uid = "me", pwd = "password")
query <- "select * from table where blah = 'blahblahblah' and blah2 = ?"

results <- function(x){
    sqlExecute(dbconn, query, x)
    sqlGetResults(dbconn)
}

ClintWeathers

It's not uncommon where I'm working to get some dataset with a needlessly longass name and needlessly longass field names.
Unfortunately, they want the field names and dataset names left as is, in situ.

Is there something like %>% for single-line stuff that would do this:

Instead of...
omgthisdatasetnameissolong$whythehelldidtheynamethisfieldlikethis <- as.character(omgthisdatasetnameissolong$whythehelldidtheynamethisfieldlikethis)

maybe this...