擅长:python、mysql、java
<pre><code>import blaze
import pandas as pd
from io import StringIO
# using StringIO for demonstration-purposes! Don't actually do this in your code!
# your code would look like: data = blaze.data('myfile.csv')
s = '''idx,MID,S,A
m.0107_f,a,1,True
m.0107_x,b,2,False
m.0107_f,c,3,True'''
data = pd.read_csv(StringIO(s))
</code></pre>
<p><code>data</code>看起来像:</p>
<pre><code> idx MID S A
0 m.0107_f a 1 True
1 m.0107_x b 2 False
2 m.0107_f c 3 True
</code></pre>
<p>现在火焰:</p>
<pre><code>csv_data = blaze.data(data)
csv_data[csv_data['idx'] == 'm.0107_f']
</code></pre>
<p>结果:</p>
<pre><code> idx MID S A
0 m.0107_f a 1 True
2 m.0107_f c 3 True
</code></pre>
<p>这里的好处是使用blaze只加载内存中满足条件的项(相当于<code>m.0107_f</code>)。你知道吗</p>