Description

This package will link multiple SQL-tables together to facilitate (potentially large-scale) analysis of data from various sources. All tables are implemented using Pandas Data.Frame objects.

Examples

In the following example, 3 tables are linked together via different columns and different data types. They form a tree: sex -> preference -> fruit. This package enables finding the favoriate fruits for ‘M’ ids by transversing this table tree automatically.

from tabletree import TableNode, TableTree, TableLink
import pandas as pd

tab1_list = list( zip([1,2,3,4], ['M', 'F', 'M', 'F'] ) )
tab2_list = list( zip(  [1,2,3,4], ['red', 'green', 'yellow', 'blue'] ) )
tab3_list = list( zip(  ['red', 'green', 'yellow', 'blue'], ['apple', 'grape', 'banana', 'berry'] ) )
tab4_list = list( zip(  [1,2,3,4], ['swimming', 'football', 'piano', 'painting'] ) )
tab1 = pd.DataFrame( tab1_list,   columns = ['tab1_v1', 'tab1_v2'])
tab2 = pd.DataFrame( tab2_list,   columns = ['tab2_v1', 'tab2_v2'])
tab3 = pd.DataFrame( tab3_list,   columns = ['tab3_v1', 'tab3_v2'])
tab4 = pd.DataFrame( tab4_list,   columns = ['tab4_v1', 'tab4_v2'])
tn1 = TableNode('sex', tab1)
tn2 = TableNode('preference', tab2)
tn3 = TableNode('fruit', tab3)
tn4 = TableNode('hobby', tab4)
ttree = TableTree(tn1)
ttree.addChildren(tn2, TableLink('tab1_v1', 'tab2_v1'))
ttree.addChildren(tn3, TableLink('tab2_v2', 'tab3_v1'))
ttree.addTN('sex', tn4, TableLink('tab1_v1', 'tab4_v1'))
print(ttree.getAncestors('hobby'))
#['sex', 'hobby']
ttree.findMatched('fruit', 'sex', tn1.df.tab1_v2 == 'M')
print( ttree.findMatched('hobby', 'sex', tn1.df.tab1_v2 == 'M') )
#tab4_v1   tab4_v2
#1         swimming
#3         piano
print(ttree)
#['sex', [['preference', [['fruit']]], ['hobby']]]

Note

Github: https://github.com/rluo/tabletree