I’m attempting to use Spyder to convert a table of temperature measurements from Fahrenheit to Celsius, but I’m getting this error: "unsupported operand type(s) for -: ‘str’ and ‘int’"

I’m using this fuction:

for f in data:
    c = round((f-32)/1.8,1)
    print(f,c)

and this is the data:

YEARMODA  TEMP   MAX   MIN  AMPLITUDE
0   20160601  65.5  73.6  54.7       18.9
1   20160602  65.8  80.8  55.0       25.8
2   20160603  68.4   NaN  55.6        NaN
3   20160604  57.5  70.9  47.3       23.6
4   20160605  51.4  58.3  43.2       15.1
5   20160606  52.2  59.7  42.8       16.9
6   20160607  56.9  65.1  45.9       19.2
7   20160608  54.2   NaN  47.5        NaN
8   20160609  49.4  54.1  45.7        8.4
9   20160610  49.5  55.9  43.0       12.9
10  20160611  54.0  62.1  41.7       20.4
11  20160612  55.4  64.2  46.0       18.2
12  20160613  58.3  68.2  47.3       20.9
13  20160614  59.7  67.8  47.8       20.0
14  20160615  63.4  70.3  49.3       21.0
15  20160616  57.8  67.5  55.6       11.9
16  20160617  60.4  70.7  55.9       14.8
17  20160618  57.3   NaN  54.0        NaN
18  20160619  56.3  59.2  54.1        5.1
19  20160620  59.3  69.1  52.2       16.9
20  20160621  62.6  71.4  50.4       21.0
21  20160622  61.7  70.2  55.4       14.8
22  20160623  60.9  67.1  54.9       12.2
23  20160624  61.1  68.9  56.7       12.2
24  20160625  65.7  75.4  57.9       17.5
25  20160626  69.6  77.7  60.3       17.4
26  20160627  60.7  70.0   NaN        NaN
27  20160628  65.4  73.0  55.8       17.2
28  20160629  65.8  73.2   NaN        NaN
29  20160630  65.7  72.7  59.2       13.5

I’m assuming the error comes from the text in the data but I’m not sure how to make it ignore the text and just try to convert the measurements.

>Solution :

You don’t need to loop. Pandas will automatically operate on all rows when you perform a calculation.

data['TEMP'] = round((data['TEMP'] - 32) / 1.8, 1)

Repeat this for all the columns containing temperatures. When the value is NaN, it will automatically return NaN as the result.