PyTables + Pandas 选择问题 [英] PyTables + Pandas Select Problems

问题描述

我有一个结构如下的 HDF5 (PyTables) 文件:

I have a HDF5 (PyTables) file structured like this:

/<User>/<API Key>
ex:
/Dan/A4N5
/Dan/B8P0
/Dave/D3Y7

每个表的结构都是这样，带有 sessionID 和存储在 epoch 中的时间:

Each table is structured like so with a sessionID and a time stored in epoch:

    sessionID        time
0   3ODE3Nzll  1467590400
1   lMGVkMDc4  1467590400
2   jNzIzNmY1  1467590400
...

我希望 Pandas 遍历每个表并获取指定日期和指定日期前一天之间的所有行.目前我有这个代码:

I want Pandas to go through each table and get all the rows between a specified date and the day before the specified date. Currently I have this code:

scriptPath = os.path.dirname(os.path.abspath(__file__))
argdate = "2016/07/14"
dayTimestamp = datetime.datetime(int(argdate[0:4]), int(argdate[5:7]), int(argdate[8:10]), tzinfo=pytz.utc)
yesterdayTimestamp = dayTimestamp - datetime.timedelta(days=1)
with pd.HDFStore(os.path.join(scriptPath, "userdatabase.h5")) as db:
    for table in db.keys():
        print(table)
        tableSplit = table.split('/')
        client = tableSplit[1]
        apiKey = tableSplit[2]
        df = db.select('{}/{}'.format(client, apiKey), where='time<=int(dayTimestamp.timestamp()) & time>=int(yesterdayTimestamp.timestamp())')
        print(df)

但是在 select 所在的行上抛出错误.

However is throwing an error on the line that select is located.

  File "tester.py", line 570, in database
    df = db.select('{}/{}'.format(client, apiKey), where='time<=int(dayTimestamp.timestamp()) & time>=int(yesterdayTimestamp.timestamp())')
  File "/usr/local/lib/python3.4/dist-packages/pandas/io/pytables.py", line 680, in select
    return it.get_result()
  File "/usr/local/lib/python3.4/dist-packages/pandas/io/pytables.py", line 1364, in get_result
    results = self.func(self.start, self.stop, where)
  File "/usr/local/lib/python3.4/dist-packages/pandas/io/pytables.py", line 673, in func
    columns=columns, **kwargs)
  File "/usr/local/lib/python3.4/dist-packages/pandas/io/pytables.py", line 4021, in read
    if not self.read_axes(where=where, **kwargs):
  File "/usr/local/lib/python3.4/dist-packages/pandas/io/pytables.py", line 3222, in read_axes
    self.selection = Selection(self, where=where, **kwargs)
  File "/usr/local/lib/python3.4/dist-packages/pandas/io/pytables.py", line 4580, in __init__
    self.terms = self.generate(where)
  File "/usr/local/lib/python3.4/dist-packages/pandas/io/pytables.py", line 4593, in generate
    return Expr(where, queryables=q, encoding=self.table.encoding)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/pytables.py", line 517, in __init__
    self.terms = self.parse()
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 727, in parse
    return self._visitor.visit(self.expr)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 311, in visit
    return visitor(node, **kwargs)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 317, in visit_Module
    return self.visit(expr, **kwargs)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 311, in visit
    return visitor(node, **kwargs)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 320, in visit_Expr
    return self.visit(node.value, **kwargs)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 311, in visit
    return visitor(node, **kwargs)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 655, in visit_BoolOp
    return reduce(visitor, operands)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 647, in visitor
    lhs = self._try_visit_binop(x)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 643, in _try_visit_binop
    return self.visit(bop)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 311, in visit
    return visitor(node, **kwargs)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 628, in visit_Compare
    return self.visit(binop)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 311, in visit
    return visitor(node, **kwargs)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 401, in visit_BinOp
    op, op_class, left, right = self._possibly_transform_eq_ne(node)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 354, in _possibly_transform_eq_ne
    right = self.visit(node.right, side='right')
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 311, in visit
    return visitor(node, **kwargs)
  File "/usr/local/lib/python3.4/dist-packages/pandas/computation/expr.py", line 615, in visit_Call_legacy
    return self.const_type(res(*args, **keywords), self.env)
TypeError: 'str' object is not callable

1) 我该如何解决这个错误?

1) How do I fix this error?

2) 是否可以遍历 HDF5 文件层次结构，以便我可以对相同的 Pandas 表进行分组?

2) Is it possible to iterate through the HDF5 file hierarchy so that I can group Pandas tables of the same ?

推荐答案

这是一个工作演示:

import io
import pandas as pd

df = pd.read_csv(io.StringIO("""
sessionID        time
3ODE3Nzll  1467590400
lMGVkMDc4  1467590400
jNzIzNmY1  1467590400
3ODE3Nzll  1467676800
lMGVkMDc4  1467676800
jNzIzNmY1  1467676800
"""), sep='\s+')

filename = 'c:/temp/aaa.h5'

store = pd.HDFStore(filename)

store.append('/aaa/df1', df, data_columns=True)
store.append('/bbb/df1', df, data_columns=True)

# let's double # of rows
df = pd.concat([df] * 2, ignore_index=True)

# and write it to HDFStore
store.append('/aaa/df2', df, data_columns=True)

print(store)

argdate = "2016/07/04"
ts_from = int(pd.to_datetime(argdate).timestamp())
ts_to = ts_from + 24*60*60

client_flt = '/aaa/'
#qry = '(time >= {0}) & (time <= {1})'.format(dayTimestamp, dayTimestamp + 24*60*60)
qry = 'time >= ts_from & time <= ts_to'
print('WHERE:\t%s' %qry)

for k in store:
    if k.startswith(client_flt):
        x = store.select(k, where=qry)
        print(k)
        print(x)

输出:

<class 'pandas.io.pytables.HDFStore'>
File path: c:/temp/aaa.h5
/aaa/df1            frame_table  (typ->appendable,nrows->6,ncols->2,indexers->[index],dc->[sessionID,time])
/aaa/df2            frame_table  (typ->appendable,nrows->12,ncols->2,indexers->[index],dc->[sessionID,time])
/bbb/df1            frame_table  (typ->appendable,nrows->6,ncols->2,indexers->[index],dc->[sessionID,time])

WHERE:  time >= ts_from & time <= ts_to
/aaa/df1
   sessionID        time
0  3ODE3Nzll  1467590400
1  lMGVkMDc4  1467590400
2  jNzIzNmY1  1467590400
/aaa/df2
   sessionID        time
0  3ODE3Nzll  1467590400
1  lMGVkMDc4  1467590400
2  jNzIzNmY1  1467590400
6  3ODE3Nzll  1467590400
7  lMGVkMDc4  1467590400
8  jNzIzNmY1  1467590400

这篇关于PyTables + Pandas 选择问题的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！