Python Indexing with Numpy Array

Question

I've generated my percentiles from this:

percentile = np.percentile(trace4['switchpoint'], [2.5, 97.5], axis=0)
print(percentile)

which returns

 array([[0.83300568, 0.46060436, 2.7449003 , 1.9982139 , 0.3987762 ,
    0.88365135, 0.22983616, 0.93329534, 2.72465413, 1.17089865,
    0.28492142, 0.6926268 , 0.66269826, 1.06656515, 0.46752127,
    1.76456569, 0.1313388 , 1.28500029, 1.20987646, 1.45013769,
    1.37307323, 0.32541516, 0.93211006, 0.4211398 , 0.96974685,
    1.43363548, 2.14078899, 1.51663305, 1.0689224 , 0.79560975,
    1.63287055, 2.09162316, 0.2881392 , 1.55702871, 1.48598385,
    3.55837488, 0.74913963, 1.04578597, 0.26495597, 0.51574263],
   [4.75729175, 4.1623318 , 7.14900756, 5.12070075, 6.92330205,
    7.00824703, 4.87916471, 6.8439851 , 5.05501153, 5.80100506,
    6.91067095, 5.57056179, 5.7841482 , 5.83522249, 6.90924864,
    8.13130971, 5.07952595, 7.49020049, 6.74093273, 6.95862249,
    6.88121063, 6.52230737, 6.84559638, 6.03093124, 7.11098186,
    6.11971399, 4.80683008, 5.82910868, 7.44122   , 7.82238211,
    5.64118377, 5.91967593, 7.18480546, 4.98052415, 5.04198767,
    4.22506871, 6.60798525, 6.4482244 , 5.92934079, 6.76150973]])

However, I would like to match it with my index as in "individual_id" which has the same length. I tried using

 pd.Series(percentile, index='individual_id')

but this doesn't work and returned error message with

 "Exception: Data must be 1-dimensional".

Can anyone help? many thanks!

Try reshaping the data pd.Series(percentile.reshape(-1)) like this — Justice_Lords
– Justice_Lords, Commented Mar 23, 2019 at 17:34
Your array appears to be 2d, with shape (2,n). The length is 2. What is this 'individual_id' supposed to be or represent? You mention 'index' but don't define it. — hpaulj
– hpaulj, Commented Mar 23, 2019 at 17:38

David Buck · Accepted Answer · 2019-03-23 21:58:00Z

2

percentile is a 2D array. Try:

pd.Series(percentile[0], index=‘individual_id’)

Some extra crap since I have to write answer of greater length than is needed to answer the question asked. This site has some silly rules for helping people solve coding problems.

edited Mar 23, 2019 at 21:58

David Buck

3,87740 gold badges54 silver badges73 bronze badges

answered Mar 23, 2019 at 17:38

Ethan

1,3831 gold badge8 silver badges8 bronze badges

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

1 Comment

Ethan Over a year ago

William is spot on. Nice catch! Depending on how you are using percentile you may want to change each array in the first dimension: pd.Series(percentile[0], index=‘individual_id’) then pd.Series(percentile[1], index=‘individual_id’) and concatenate them OR follow Justice_Lords advice and reshape the array: percentile.reshape(-1)

William Scott · Accepted Answer · 2019-03-23 19:13:41Z

0

I agree with Ehtan's answer to solve your error.

but, by just taking the data on percentile[0], you will lose the percentile data from trace4['switchpoint'] on your second value in [2.5, 97.5].

So instead, try changing [2.5, 97.5] to a single value or just take an average. analyse why you need that 2.5 or 97.5.

so you can try something like this:

percentile = np.percentile(trace4['switchpoint'], 50, axis=0)

pd.Series(percentile, index='individual_id')

I'm not saying what you did is wrong. But it's meaningless to generate for two values and use one.

Hope it helps ;)

answered Mar 23, 2019 at 19:13

William Scott

3191 gold badge5 silver badges14 bronze badges

Collectives™ on Stack Overflow

Python Indexing with Numpy Array

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related