如何选择要在熊猫中绘图的DataFrame列？

Question

I have a python DataFrame from Pandas

waitEvent   instance    snapDate    gc cr block 3-way   gc current block 3-way  log file sync

Every time I construct this DataFrame, the number and name of columns after the snapDate are different.

I would need to construct a new DataFrame by selecting snapDate as the indexcolumn but the remaining columns should be dynamically selected for plotting any ideas. How can I achieve this?

for the df above, I should always select the snapDate column and all the columns after the snapDate column

The number and name of the columns after the snapDate column will vary.

My objective is to do

df[snapDate,col1,col2,col3].plot()
df[snapDate,col7,col8].plot()
..

I want to create a plot from the DataFrame by always picking the SnapDate column and the remaining columns which depending on the DataFrame could be 2 or 3 or 4 etc.

Answer 1

You can use get_loc to get the index position of the column of interest and then use this to index your df:

In [374]:
cols = ['waitEvent','instance','snapDate','gc cr block 3-way','gc current block 3-way','log file sync']
df = pd.DataFrame(columns = cols)
df

Out[374]:
Empty DataFrame
Columns: [waitEvent, instance, snapDate, gc cr block 3-way, gc current block 3-way, log file sync]
Index: []

In [378]:    
snapDateIdx = df.columns.get_loc('snapDate')
snapDateIdx

Out[378]:
2

In [379]:
df.ix[:,snapDateIdx:]

Out[379]:
Empty DataFrame
Columns: [snapDate, gc cr block 3-way, gc current block 3-way, log file sync]
Index: []