R to Python subsetting via vector
Question:
I’m a python newbie but have some R experience. In R if I’d like to subset a data.frame I can use a variable to do something like this:
# Columns
# Assign column names to variable
colsToUse <- c('col1','col2','col3')
# Use variable to subset
df2 <- df1[,colsToUse]
# Rows
# Assign column names to variable
rowsToUse <- sample(1:nrows(df1), 500)
# Use variable to subset
df3 <- df1[rowsToUse,]
How would I do this in python?
Answers:
Based on your stated use of pandas
colsToUse = ['col1', 'col2', 'col3']
rowsToUse = np.random.choice(range(len(df1)), 500)
df2 = df1.ix[:, colsToUse]
df3 = df1.ix[rowsToUse, :]
There are also some other DataFrame helper functions for indexing: df1.loc, df1.iloc, and df1.xs.
It’s also helpful to look at the guide NumPy for MATLAB Users which also often answers questions for R users too, at least when dealing with just a numpy.ndarray).
I’m a python newbie but have some R experience. In R if I’d like to subset a data.frame I can use a variable to do something like this:
# Columns
# Assign column names to variable
colsToUse <- c('col1','col2','col3')
# Use variable to subset
df2 <- df1[,colsToUse]
# Rows
# Assign column names to variable
rowsToUse <- sample(1:nrows(df1), 500)
# Use variable to subset
df3 <- df1[rowsToUse,]
How would I do this in python?
Based on your stated use of pandas
colsToUse = ['col1', 'col2', 'col3']
rowsToUse = np.random.choice(range(len(df1)), 500)
df2 = df1.ix[:, colsToUse]
df3 = df1.ix[rowsToUse, :]
There are also some other DataFrame helper functions for indexing: df1.loc, df1.iloc, and df1.xs.
It’s also helpful to look at the guide NumPy for MATLAB Users which also often answers questions for R users too, at least when dealing with just a numpy.ndarray).