Numpy delete multiple rows matching criteria

Question

I have a numpy array of folowing structure

sb = np.genfromtxt(open('HomePage/TodayList.txt', 'rb'),
                   delimiter=',', skiprows=0,
                   dtype=[('DataBase', np.str_, 16), ('Mode', np.str_, 16),
                          ('SMB', np.str_, 16),('Desc', np.str_, 128), 
                          ('Res', np.str_, 16), ('RightCnt', np.float64), 
                          ('PercentCnt', np.float64), ('ModelType', np.float64)])

The 6th column 'PercentCnt' which can be accessed by name 'PercentCnt' contains numbers from 0 to 50 the 7th column 'ModelType' contains numbers from 0 to 5 so i need to remove or delete array rows which match these criteria 'PercentCnt'<50 and 'ModelType'<2.

Ami Tavory · Accepted Answer · 2016-09-20 08:19:24Z

2

The condition

sb['PercentCnt'] >= 50

is the condition for keeping things on this column, and the condition

sb['ModelType'] >= 2

is the same for the other column.

You can combine these with np.logical_and:

keep = np.logical_and(sb['PercentCnt'] >= 50, sb['ModelType'] >= 2)

Finally, just keep the rows you wish to keep:

sb[keep]

answered Sep 20, 2016 at 8:19

Ami Tavory

76.7k13 gold badges152 silver badges196 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

jotasi · Accepted Answer · 2016-09-20 08:18:29Z

1

You can find all rows matching your criteria by use of a column-wise comparison for PercentCnt and ModelType and connection using np.logical_and. Doing that, you actually copy all other rows rather than to delete the ones you wanted to get rid of, but the effect is the same.

sb = sb[np.logical_and(sb["PercentCnt"]>=50, sb["ModelType"]>=2)]

answered Sep 20, 2016 at 8:18

jotasi

5,2072 gold badges31 silver badges53 bronze badges

Collectives™ on Stack Overflow

Numpy delete multiple rows matching criteria

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related