python - Handling HUGE numbers in numpy or pandas -

i doing competition provided data anonymized. quite few of columns have huge values. largest 40 digits long! used pd.read_csv columns have been converted objects result.

my original plan scale data down since seen objects can't arithmetic on these.

does have suggestion on how handle huge numbers in pandas or numpy?

note i've tried converting value uint64 no luck. error "long big convert"

you can use pandas converters call int or other custom converter function on string being imported:

import pandas pd  stringio import stringio  txt='''\ line,big_num,text 1,1234567890123456789012345678901234567890,"that sure big number" 2,9999999999999999999999999999999999999999,"that bigger number" 3,1,"tiny" 4,-9999999999999999999999999999999999999999,"really negative" '''  df=pd.read_csv(stringio(txt), converters={'big_num':int})  print df 

prints:

   line                                    big_num                           text 0     1   1234567890123456789012345678901234567890      sure big number 1     2   9999999999999999999999999999999999999999  bigger number 2     3                                          1                           tiny 3     4  -9999999999999999999999999999999999999999                negative 

now test arithmetic:

n=df["big_num"][1] print n,n+1  

prints:

9999999999999999999999999999999999999999 10000000000000000000000000000000000000000 

if have values in column might cause int croak, can this:

txt='''\ line,big_num,text 1,1234567890123456789012345678901234567890,"that sure big number" 2,9999999999999999999999999999999999999999,"that bigger number" 3,0.000000000000000001,"tiny" 4,"a string","use 0 strings" '''  def conv(s):     try:         return int(s)     except valueerror:         try:             return float(s)         except valueerror:             return 0          df=pd.read_csv(stringio(txt), converters={'big_num':conv}) print df 

prints:

   line                                   big_num                           text 0     1  1234567890123456789012345678901234567890      sure big number 1     2  9999999999999999999999999999999999999999  bigger number 2     3                                     1e-18                           tiny 3     4                                         0              use 0 strings 

then every value in column either python int or float , support arithmetic.