Sort by p-value

[plumdeq]

Thanks for a fantastic library. Do you know if it is possible to sort by the p-value? Currently I have to write the whole table to csv, then load it as a Pandas DataFrame, and then sort it. Sorting the rows with categorial variables is, however, not possible in this case...

[tompollard]

Thanks @plumdeq :) We'll try to look at p-value sorting and #83 in the next few days.

[tompollard]

Not got around to this yet (sorry), but one thing to note is that the pd.DataFrame object can be accessed with the tableone attribute (e.g. mytable.tableone), so there shouldn't be a need to do your custom sort by writing to CSV first.

[plumdeq]

(Double) Thank you, Tom!

[tompollard]

Version 0.6.6 allows you to sort by the following columns:

• 'Missing'
• 'P-Value'
• 'P-Value (adjusted)'
• 'Test'

Hope this helps, and give me a shout if you hit any issues! Example below:

# import libraries
from tableone import TableOne
import pandas as pd

# load sample data into a pandas dataframe
url="https://raw.githubusercontent.com/tompollard/tableone/master/data/pn2012_demo.csv"
data=pd.read_csv(url)

# columns to summarize
columns = ['SysABP', 'Height', 'Weight', 'ICU']

# columns containing categorical variables
categorical = ['ICU']

# create tableone with the input arguments
mytable = TableOne(data, columns=columns, categorical=categorical, 
                   label_suffix=True, groupby = "death", pval=True,
                   sort="P-Value")

print(mytable.tabulate(tablefmt = "github"))

		Missing	0	1	P-Value	Test
n			864	136
ICU, n (%)	CCU	0	137 (15.9)	25 (18.4)	<0.001	Chi-squared
	CSRU		194 (22.5)	8 (5.9)
	MICU		318 (36.8)	62 (45.6)
	SICU		215 (24.9)	41 (30.1)
SysABP, mean (SD)		291	115.4 (38.3)	107.6 (49.4)	0.134	Two Sample T-test
Height, mean (SD)		475	170.3 (23.2)	168.5 (11.3)	0.304	Two Sample T-test
Weight, mean (SD)		302	83.0 (23.6)	82.3 (25.4)	0.782	Two Sample T-test