Extract data from string in python

Question

I have a .csv file with data constructed as [datetime, "(data1, data2)"] in rows and I have managed to import the data into python as time and temp, the problem I am facing is how do I seperate the temp string into two new_temp columns in float format to use for plotting later on?

My code so far is:

import csv
import matplotlib.dates as dates

def getColumn(filename, column):
    results = csv.reader(open('logfile.csv'), delimiter = ",")
    return [result[column] for result in results]
time = getColumn("logfile.csv",0)
temp = getColumn("logfile.csv",1)
new_time = dates.datestr2num(time)
new_temp = [???]

When I print temp I get ['(0.0, 0.0)', '(64.4164, 66.2503)', '(63.4768, 65.4108)', '(62.7148, 64.6278)', '(62.0408, 63.9625)', '(61.456, 63.2638)', '(61.0234, 62.837)', '(60.6823, 62.317)',...etc]

If anyone can help me then thanks in advance.

Possible duplicate of [Parse a tuple from a string?](http://stackoverflow.com/questions/9763116/parse-a-tuple-from-a-string) — Moon Cheesez, Jul 11 '16 at 12:49
As a side note, you should read the CSV only once and then extract the columns. Currently you read the whole file over and over. — CodeMonkey, Jul 11 '16 at 12:52
I checket that solution and wrote make_tuple(temp) and get the error: ValueError: malformed string. And i also tried the tuple(int(x) for x in temp[1:-1].split(',')) and get the error: AttributeError: 'list' object has no attribute 'split' — Gísli Trausti Jóhannesson, Jul 11 '16 at 13:06
when i change the "string" into my "temp" variable then i get the error: TypeError: expected string or buffer — Gísli Trausti Jóhannesson, Jul 11 '16 at 13:12
As an extra info then here is a drive [link](https://drive.google.com/open?id=0B-yaioMxfCgyVEhwOTBiX2JNbUk) to the logfile then you could try running the solution with the given code. — Gísli Trausti Jóhannesson, Jul 11 '16 at 13:15
You can try to replace string with str(temp). Maybe this is enough to make it work. — Ohumeronen, Jul 11 '16 at 13:23

score 1 · Answer 1 · edited May 23 '17 at 12:22

1

Going from the answer in Parse a tuple from a string?

from ast import literal_eval as make_tuple
temp = ['(0.0, 0.0)', '(64.4164, 66.2503)', '(63.4768, 65.4108)', '(62.7148, 64.6278)', '(62.0408, 63.9625)', '(61.456, 63.2638)', '(61.0234, 62.837)', '(60.6823, 62.317)']

tuples = [make_tuple(stringtuple) for stringtuple in temp]

and you have an array of tuples of doubles.

I undeleted the post and made it a full answer, because apparently it wasn't clear enough to reference the other post.

edited May 23 '17 at 12:22

Community

1
1

answered Jul 11 '16 at 12:47

CodeMonkey

4,067
1
31
43

If it has already been answered, post it as a comment, not an answer. E.g. Possible duplicate of http://stackoverflow.com/questions/9763116/parse-a-tuple-from-a-string – Moon Cheesez Jul 11 '16 at 12:47
1

@user3387223: Thanks for sharing this. – Ohumeronen Jul 12 '16 at 11:48

Ohumeronen · Accepted Answer · 2016-07-12T09:17:02.743

1

You may use this code:

import re
string = "['(0.0, 0.0)', '(64.4164, 66.2503)', '(63.4768, 65.4108)', '(62.7148, 64.6278)', '(62.0408, 63.9625)', '(61.456, 63.2638)', '(61.0234, 62.837)', '(60.6823, 62.317)']"
data = re.findall('[+-]?\d+\.\d+e?[+-]?\d*', string)
data = zip(data[0::2], data[1::2])
print [float(d[0]) for d in data]
print [float(d[1]) for d in data]

edited Jul 12 '16 at 09:17

answered Jul 11 '16 at 12:52

Ohumeronen

1,769
2
14
28

1

This will fail if you have some differently formatted floats (e.g. 5.2456e10). Better use Python's built in parser for evaluating Python expressions in strings as suggested in the comments. – CodeMonkey Jul 11 '16 at 12:54
No special format was mentioned but thanks for the suggestion. – Ohumeronen Jul 11 '16 at 12:55
Does it work with 5.2456e-10, 5.2456e+9, -5.2456e-10, .1, 1.? – CodeMonkey Jul 11 '16 at 13:07
Yes, now it does, but I am sure you will find another case ;-) Deutsche Korrektheit was? – Ohumeronen Jul 11 '16 at 13:09
Signs can also be leading? Ja, genau :-D! – CodeMonkey Jul 11 '16 at 13:10
Ok use a parser if this does not suit your needs. – Ohumeronen Jul 11 '16 at 13:11
1

I found that the suggestion of user3387223 was quite important so I improved the solution again to allow negative values. – Ohumeronen Jul 12 '16 at 09:18

Extract data from string in python

2 Answers2