Python + SQL in RStudio Notebooks

database.Rmd

---
title: "dplyr Database Example"
output: github_document
---

```{r}
library(DBI)
library(RSQLite)

# Can be any DBI compliant database connection
con <- RSQLite::datasetsDb()
RSQLite::dbListTables(con)
```



```{sql connection=con}
Select * FROM trees
```

```{sql connection=con, output.var='trees'}
Select * FROM trees
```

```{r}
trees
```

experimental_python.Rmd

---
title: "Interacting with Python from R"
output: html_notebook
---

# Using python from within R

The following uses Bryan Lewis's `python` package. All of the code *is R code*, where R is acting as a wrapper that calls python objects. One thing that is handy about this is the python objects can stick around!

## Setup

```{r}
#devtools::install_github("bwlewis/python")
library(python)
np = import("numpy")
```

## Go back and forth between R any Python objects.

From R to Python:

```{r}
set.seed(1)
x = matrix(rnorm(9), nrow=3)
pyX = np$array(x)
print(pyX)
```

Note that the `p` variable is really a Python object, for example the output
above was printed by Python.

```{r}
class(pyX)
```

From python object back into R:

```{r}
Rx = R(pyX)
class(Rx)
identical(Rx, x)
```

## Python methods

Use the `pyhelp()` function if you want help on something. At the moment,
`pyhelp()` needs you to provide a full valid Python name (with dots). For instance:

```{r}
pyhelp("numpy.ndarray.prod")
```


```{r}
pyX$prod() 
```

python.Rmd

---
title: "R & Python Code Chunks"
output:
  html_document: default
  html_notebook: default
---


Comparing pricipal component analysis in R and Python.

## R

```{r}
data(iris)
iris.data <- iris[, 1:4]
pca <- princomp(iris.data)
scores <- pca$scores[, 1:2]
scores[1:6, ]
```


## Python

```{python}
from sklearn import decomposition, datasets
iris = datasets.load_iris()
pca = decomposition.PCA(n_components=2)
pca.fit(iris.data)
scores = pca.transform(iris.data)
print scores[0:6, ]
```


You can run multiple versions:

## Python Versions

```{python, engine.path="/usr/bin/python2.6"}
import sys
print(sys.version)
```

```{python, engine.path="/usr/bin/python2.7"}
import sys
print(sys.version)
```