pandas使用问题 - Githubissues

dengdaiyemanren commented 8 years ago

http://stackoverflow.com/questions/34383000/pandas-to-sql-all-columns-as-nvarchar

dtype = {'date':sqlalchemy.DateTime(),'code':sqlalchemy.Integer} df.to_sql('hist_stock_data',engine,index=True,if_exists="append",dtype= dtype)

dengdaiyemanren commented 8 years ago

cursors.py:206: Warning: Incorrect string value

http://stackoverflow.com/questions/2108824/mysql-incorrect-string-value-error-when-save-unicode-string-in-django

dengdaiyemanren commented 8 years ago

乱码问题解决 http://fc-lamp.blog.163.com/blog/static/174566687201210285137981/

dengdaiyemanren commented 8 years ago

地址： http://pandas.pydata.org/pandas-docs/version/0.18.1/10min.html

dengdaiyemanren commented 8 years ago

pandas 与 DataFrame http://www.aboutyun.com/thread-12358-1-1.html

dengdaiyemanren commented 8 years ago

##
ts11 = pd.Series([10.1,10.5,10.2,10.10,12,10], index=['2010-01-01','2010-01-02','2010-01-03','2010-01-04','2010-01-05','2010-01-06'])
ts11 = ts11.cumsum()
ts11.plot()

## a dict 
ts12 = {'a' : 0., 'b' : 1., 'c' : 2.}
ts13 = pd.Series(ts12)
ts13.plot()

##From dict of Series or dicts

d = {'code1' : pd.Series([2., 3., 4.], index=['a', 'b', 'c']),
    'code2' :  pd.Series([1., 2., 3., 4.], index=['a', 'b', 'c', 'd'])}
df = pd.DataFrame(d)  
df.plot()
#print df 

## ndarrays / lists
d2 = {'one' : [1., 2., 3., 4.], 'two' : [4., 3., 2., 1.]}
df2 = pd.DataFrame(d2) 
df2 = pd.DataFrame(d2, index=['a', 'b', 'c', 'd'])

#df2.plot()

## From a list of dicts
data3 = [{'a': 1, 'b': 2}, {'a': 5, 'b': 10, 'c': 20}]
df3 = pd.DataFrame(data3) 
#df3.plot()

## finish
ts11 = pd.Series([10.1,10.5,10.2,10.10,12,10], index=['2010-01-01','2010-01-02','2010-01-03','2010-01-04','2010-01-05','2010-01-06'])

d = {'000033' : pd.Series([10.1,10.5,10.2,10.10,12,10], index=['2010-01-01','2010-01-02','2010-01-03','2010-01-04','2010-01-05','2010-01-06']),
    '000034' :  pd.Series([4,5,5.2,5.4,6,5.5], index=['2010-01-01','2010-01-02','2010-01-03','2010-01-04','2010-01-05','2010-01-06'])}
df = pd.DataFrame(d)  

#ts11 = ts11.cumsum()
#ts11.plot()

dengdaiyemanren commented 7 years ago

DataFrame

http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.truncate.html

DataFrame可视化

http://pandas.pydata.org/pandas-docs/version/0.13.1/visualization.html

dataframe 合并contact

http://pandas.pydata.org/pandas-docs/stable/merging.html

同一个坐标展示2个dataframe

误区：前面想着把datframe 通过同名覆盖，然后加个z坐标区分，这个难度很高，很难实现方法：把dataframe的列生成新的列，在一个dataframe里展示是可以的。

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
   df = pd.DataFrame(np.random.randn(10,4),
 index=pd.date_range('20100101',periods=10), columns=list('ABCD'))

    df = dfcontact
    df = df.cumsum()
    plt.figure()
    df.plot()
    plt.legend(loc='best')

from chartpy import Chart, Style, Canvas
import tushare.stock.trading as fd

    df1 = fd.get_hist_data("300431",'2016-03-12')
    df1['code'] ="300431"
    df1['300431'] = df1['close']
    df2 = fd.get_hist_data("000786",'2016-03-12')
    df2['000786'] = df2['close']

    framses =[df1.filter(items=['300431']),df2.filter(items=['000786'])]
    #dfcontact = pd.concat(framses)
    dfcontact = df1.filter(items=['300431']).append(df2.filter(items=['000786']))

    chart_plotly1 = Chart(df=dfcontact, chart_type='line', engine='plotly',
                         style=Style(title="股价对比图", source="Quandl/Fred", scale_factor=-2, width=500, height=300, silent_display=True))

    text = "A demo of chartpy canvas!!"

    # using plain template
    canvas = Canvas([[chart_plotly1]])

    canvas.generate_canvas(silent_display=False, canvas_plotter='plain')

dengdaiyemanren commented 7 years ago

apply

def sumHostAndGust(row,col1,col2):
        return row[col1] + row[col2]
    dff = dfn.apply(sumHostAndGust,axis=1,col1='hostScore', col2='awayScore')

dengdaiyemanren commented 7 years ago