python - Why does "numpy.mean" return 'inf'? -

python - Why does "numpy.mean" return 'inf'? -

i need calculate mean in columns of array more 1000 rows.

np.mean(some_array) gives me inf output

but pretty sure values ok. loading csv here data variable , column 'cement' "healthy" point of view.

in[254]:np.mean(data[:230]['cement']) out[254]:275.75

but if increment number of rows problem starts:

in [259]:np.mean(data[:237]['cement']) out[259]:inf

but when @ data

in [261]:data[230:237]['cement'] out[261]: array([[ 425. ], [ 333. ], [ 250.25], [ 491. ], [ 160. ], [ 229.75], [ 338. ]], dtype=float16)

i not find reason behaviour p.s happens in python 3.x using wakari (cloud based ipython)

numpy version '1.8.1'

i loading info with:

no_col=9 conv = lambda valstr: float(valstr.replace(',','.')) c={} in range(0,no_col,1): c[i] = conv data=np.genfromtxt(get_data,dtype=float16 , delimiter='\t', skip_header=0, names=true, converters=c)

i guess problem precision (as others have commented). quoting straight documentation mean() see

notes

the arithmetic mean sum of elements along axis divided number of elements.

note floating-point input, mean computed using same precision input has. depending on input data, can cause results inaccurate, float32 (see illustration below). specifying higher-precision accumulator using dtype keyword can alleviate issue.

since array of type float16 have limited precision. using dtype=np.float64 alleviate overflow. see examples in mean() documentation.

python numpy