Search the maximum value in a list of Series: ValueError: The truth value of a Series is ambiguous

Question

I need to find the maximum value among values in a list. Every element of this list is a pd.Series so when I use the function max(list) I get this error:

"ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all()."

How can I avoid this problem?

From a Dataframe that look like this:

             Date                Actual 
0     2016-12-30 12:30:00          1800
1     2016-12-30 13:00:00          1800
2     2016-12-30 13:30:00          1600 
.
.
.
1256   201-05-30 13:30:00          1500

I create my list:

for single_date in datetime_range(start_date, end_date, timedelta(minutes=30)):
        b = find_nearest(df['Date'], single_date)
        row = df4[df4['Date'] == b ].index.tolist()  
        en=(df4.iloc[row]['Actual (kW)'])
        energy.append(en)

I need to find the maximum value in the list energy. Thank you

Welcome to StackOverflow. Please take the time to read this post on [how to provide a great pandas example](http://stackoverflow.com/questions/20109391/how-to-make-good-reproducible-pandas-examples) as well as how to provide a [minimal, complete, and verifiable example](http://stackoverflow.com/help/mcve) and revise your question accordingly. These tips on [how to ask a good question](http://stackoverflow.com/help/how-to-ask) may also be useful. — jezrael, Apr 25 '18 at 11:27
Sorry, I am basically a beginner. Every element of my list is taken from a Dataframe look like this: ` Date Actual (kW) 0 2016-12-30 12:30:00 180 1 2016-12-30 13:00:00 180 2 2016-12-30 13:30:00 180` — Arianna Sorrentino, Apr 25 '18 at 11:34
@shivsn I obtain < built-in method max of numpy.ndarray object at 0x000001FEAD6655D0 > — Arianna Sorrentino, Apr 25 '18 at 11:40
@AriannaSorrentino you will have to post a sample and desired output. — shivsn, Apr 25 '18 at 11:42
@AriannaSorrentino It is not clear what "the maximum value among values in a list" is when the list items are `Series`. — Stop harming Monica, Apr 25 '18 at 11:45
I obtain `code`dataframe nan C:\Users\arianna\Anaconda3\lib\site-packages\numpy\core\_methods.py:26: RuntimeWarning: invalid value encountered in reduce return umr_maximum(a, axis, None, out, keepdims)`code` — Arianna Sorrentino, Apr 25 '18 at 12:09

score 0 · Answer 1 · answered Apr 25 '18 at 13:12

In order to calculate the maximum of a list, Python needs to compare the elements within the list. In your case, each element of the list is a pandas Series. And the error comes from the fact that there is no single True/False answer in the comparison of two series (as the comparison is done element by element).

How do you want to determine if one pandas Series is bigger than the other?

If I assume that each Series has only one value, you can do:

max(list_name, key=lambda x: x.values[0])

In general, in key you specify a function that is applied to the elements of the list before the comparison, hence here I passed a function that takes the first (and only) value from each Series.

Oliver Baumann · Accepted Answer · 2018-04-25T14:08:20.600

0

Since you are working with pandas Dataframes, you can use these objects to find the maximum value...

... of the whole frame, i.e. the column-wise maxima; in a frame with m columns you'd get m maximal values
... of individual columns

For both, see here as reference.

So in your case, you should be able to simply do:

df['Actual'].max()

where df is the dataframe from your example.

If you have to filter down to a range of values, that is possible too. Consider this (arbitrary) example for the syntax:

df[df['Foo'] >= 3]['Bar'].max()

This will take column 'Foo', select everything where 'Foo' is larger-or-equal to 3, and then take column 'Bar' and print the maximum for only that filtered range of 'Bar'.

I suggest you take a look at how selection is done in Pandas, there's some neat examples there!

edited Apr 25 '18 at 14:08

answered Apr 25 '18 at 13:48

Oliver Baumann

2,209
1
10
26

But using that command will I obtain the maximum of the entire column? because I need tha maximum in a slice of the Dataframe column. – Arianna Sorrentino Apr 25 '18 at 14:01
You can filter down based on other values. I edited my anser with a rather random example, does that help you? – Oliver Baumann Apr 25 '18 at 14:08
Actually I already filter the Dataframe. I try to modify my code to exploit yor suggestion. Thanks – Arianna Sorrentino Apr 25 '18 at 14:48
@AriannaSorrentino, did you make any progress on your problem? – Oliver Baumann Apr 27 '18 at 10:42
yes @oliver the solution you suggested was perfect! Thank you :) – Arianna Sorrentino Apr 28 '18 at 15:48
@AriannaSorrentino that's great news :) if you don't mind, please accept this answer by clicking on the check mark, this will help others navigate the space (and reward me with reputation points ;-) ) – Oliver Baumann Apr 28 '18 at 19:27

Search the maximum value in a list of Series: ValueError: The truth value of a Series is ambiguous

2 Answers2