将ndarray保存到文本文件的麻烦

Question

I have a set of data in the format as below:

name q1 q2 q3
ER_1 0 0 0
ER_2 0 0 0

I have written a code that does a calculation by reading in this data table and then if the particular calculation concerns one of the names mentioned ex: ER_1 then the q1, q2 and q3 will get updated. I am able to do this and modify the table except that I have trouble saving it in the same format. Here is my attempt which results in a txt file with all elements in a single line ie : name,q1,q2,q3,ER_1,0,0,0,ER_2,0,0,0

data=numpy.loadtxt('table.txt', dtype=str,usecols=(0,1,2,3,4))#,unpack=True)

NAME='ER_1'

lst=[0,1,2]

for i in lst:
    if NAME in data[i][0]:
        data[i][1] = '27'


data.tofile('newdata.txt',sep=',')

code as requested:

so you'll need to make a text file called - massout.txt which looks like this:

Source   M_ratio  R_ratio   B_ratio   F_tot F_red   F_blue  c3
ER_1     0           0      0         0       0        0     0 
ER_2     0           0      0         0       0        0     0     

then just copy and run this:

massout=numpy.loadtxt('massout.txt', dtype=str,usecols=(0,1,2,3,4,5,6,7))#,unpack=True)
source='ER_1'

lst=[0,1,2]

for i in lst:
    if source in massout[i][0]:
        massout[i][1] = '27'


numpy.savetxt('newdata.txt', massout, delimiter='', fmt='%s\n'  )

Answer 1

The tofile method writes the data "flattened". It writes the elements sequentially to the file, losing any shape information about the array.

Your use of the function numpy.savetxt needs a couple tweaks. To separate the fields with spaces, use delimiter=' ' (that's a space character, not an empty string). Don't include '\\n' in the fmt argument. The following works for me:

numpy.savetxt('newdata.txt', massout, delimiter=' ', fmt='%s')

---

PS With a 2-d numpy array, it is more efficient and stylistically preferable to index the array as massout[i, j] instead of massout[i][j] . That is, you should write

    massout[i, 1] = '27'

instead of

    massout[i][1] = '27'