How do I get the 10 smallest numbers of an array?-CodePudding

This is my code:

from astropy.io import fits
import pandas
import matplotlib.pyplot as plt
import numpy as np
import heapq 

datos = fits.open('/home/citlali/Descargas/Lista.fits')
data = datos[1].data

#Linea [SIII] 9532
Mask_1 = data['flux_[SIII]9531.1_Re_fit'] / data['e_flux_[SIII]9531.1_Re_fit'] > 5
newdata1 = data[Mask_1]

H1_alpha = newdata1['log_NII_Ha_Re']

H1_beta = newdata1['log_OIII_Hb_Re']

M = H1_alpha < -0.9

newx = H1_alpha[M] #This is my array where I need the smallest 10 numbers
newy = H1_beta[M]  

sm = heapq.nsmallest(10, newx)

plt.plot(sm, newy, 'ro')

I want the 10 smallest numbers of newx but also I need "y" values (newy) of this numbers and idk how to get them. Thanks.

CodePudding user response：

the documentation for heapq.nsmallest shows that you can give it a key:

heapq.nsmallest(n, iterable, key=None)

this means that you can zip the newx and newy values together and then choose the nsmallest based on the newx values.

M = H1_alpha < -0.9

newx = H1_alpha[M]
newy = H1_beta[M]  

sm = heapq.nsmallest(10, zip(newx, newy), key= lambda x: x[0])

plt.plot([i[0] for i in sm], [i[1] for i in sm], 'ro')

CodePudding user response：

Combine newx and newy into a list of tuples. Then you can get the 10 smallest from this, and split them back into separate lists to sort.

sm = heapq.nsmallest(10, zip(newx, newy)) # zip them to sort together
newx, newy = zip(*sm) # unzip them
plt.plot(newx, newy, 'ro')

CodePudding user response：

If your data is a basic Python data structure, you can use zip as user @barmar indicated.

import heapq

newx = [10, 20, 30, 40, 50, 60, 70, 80]
newy = [11, 21, 31, 41, 51, 61, 71, 81]

sm = heapq.nsmallest(4, zip(newx, newy))
print(sm)

If you're using pandas, you can use the .nsmallest() method of the series and use the index of the result to get the matching results from the other series (saving you the creation of a 3rd data structure that has a copy of all the data again):

from pandas import Series

newx = Series(newx)
newy = Series(newy)

smx = newx.nsmallest(4)
smy = newy[smx.index]
print(smx, smy)

If you're just using numpy, this works:

import numpy as np

anewx = np.array(newx)
anewy = np.array(newy)

smxi = np.argpartition(anewx, 4)[:4]
print(anewx[smxi])
print(anewy[smxi])

Results of the combined code running:

[(10, 11), (20, 21), (30, 31), (40, 41)]
0    10
1    20
2    30
3    40
dtype: int64 0    11
1    21
2    31
3    41
dtype: int64
[20 10 30 40]
[21 11 31 41]

CodePudding user response：

def get(i, lis):
    lis.sort()
    log = []
    for op in range(i):
        log.append(op)
    return log

mylist = [5, 6, 8, 9, 1, 14, 52, 10, 65, 12, 65]
print(get(10, mylist))