如何报告列意味着数据表过滤 Python Dash
How to report column means with datatable filtering Python Dash
我正在使用 Python Dash 构建一个应用程序，我可以在其中显示破折号数据表，并将数据表（基础 pandas 数据框）的列平均值报告为破折号数据表在它下面..到目前为止，我已经尝试过：
import dash
from dash.dependencies import Input, Output
import dash_table
import dash_core_components as dcc
import dash_html_components as html
import pandas as pd
import json
wide_data = pd.DataFrame(np.array([[1.24, 2.34, 3.234], [4.24, .45, .06], [7, 8, 9]]),
columns=['a', 'b', 'c'])
df = pd.DataFrame(wide_data)
df = pd.DataFrame(wide_data)
df_floats = df.select_dtypes(include=['float64'])
df_floats_means = df_floats.mean(numeric_only=True)
df_floats_means = pd.DataFrame(df_floats_means)
df_floats_means_T = df_floats_means.T
app = dash.Dash(__name__)
app.layout = html.Div([
html.H1('Customer Complaints Dashboard'),
html.Label('Read and Writing Queries Filtering'),
# html.Label2('If "Read filter_query" is chosen, use operators (le, ge, <, <=, >, >=) + tag value to filter'),
# html.Label3('If "Write filter_query" is chosen, ___'),
dcc.RadioItems(
id='filter-query-read-write',
options=[
{'label': 'Read filter_query', 'value': 'read'},
],
value='read'
),
html.Br(),
dcc.Input(id='filter-query-input', placeholder='Enter filter query'), #this is for write functionality
html.Div(id='filter-query-output'), #this is associated with read screen
html.Hr(),
dash_table.DataTable(
id='datatable-advanced-filtering',
columns=[
{'name': i, 'id': i, 'deletable': True} for i in df.columns
# omit the id column
if i != 'id'
],
data=df.to_dict('records'),
editable=True,
page_action='native',
page_size=10,
filter_action="native",
fixed_columns={'headers': True, 'data': 1},
style_table={'minWidth': '100%'}
),
html.Hr(),
html.Div(id='datatable-query-structure', style={'whitespace': 'pre'}),
html.Hr(),
html.Label('Dataframe Means'),
dash_table.DataTable(
id='datatable-advanced-filtering2',
columns=[
{'name': i, 'id': i, 'deletable': True} for i in df_floats_means_T.columns
# omit the id column
if i != 'id'
],
data=df_floats_means_T.to_dict('records'),
editable=True,
page_action='native',
page_size=10,
# filter_action="native",
fixed_columns={'headers': True, 'data': 1},
style_table={'minWidth': '100%'}
),
html.Hr(),
html.Div(id='datatable-query-structure2', style={'whitespace': 'pre'}),
html.Hr(),
])
@app.callback(
[Output('filter-query-input', 'style'),
Output('filter-query-output', 'style')],
[Input('filter-query-read-write', 'value')]
)
def query_input_output(val):
input_style = {'width': '100%'}
output_style = {}
if val == 'read':
input_style.update(display='none')
output_style.update(display='inline-block')
else:
input_style.update(display='inline-block')
output_style.update(display='none')
return input_style, output_style
@app.callback(
Output('filter-query-output', 'children'),
[Input('datatable-advanced-filtering', 'filter_query')]
)
def read_query(query):
if query is None:
return "No filter query"
return dcc.Markdown('`filter_query = "{}"`'.format(query))
@app.callback(
Output('datatable-advanced-filtering2', 'data'),
[Input('datatable-advanced-filtering', 'data')]
)
def update_means(data):
dff = df
df_floats = dff.select_dtypes(include=['float64'])
df_floats_means = df_floats.mean(numeric_only=True)
df_floats_means = pd.DataFrame(df_floats_means)
df_floats_means_T = df_floats_means.T
converted_means = df_floats_means_T.to_dict('records')
return([converted_means])
if __name__ == '__main__':
app.run_server(threaded=True)
但我知道这是不对的，因为我在应用程序布局的这一部分进行任何过滤之前计算均值 id='datatable-advanced-filtering2'。大家能帮我回调一下吗？
我 100% 是 dash 的新手，而且我还没有看到很多关于如何正确执行此操作的指南。任何帮助将不胜感激。谢谢。
我相信您可以使用主 DataTable 的 data 属性 作为回调函数的输入，并将输出设为 data 属性你的意思是数据表。其结构可能类似于：
@app.callback(
Output('datatable-advanced-filtering2', 'data'),
[Input('datatable-advanced-filtering', 'data')]
)
def update_means(data):
# Use data to construct dataframe
# Find means using same method used at top of post
# Convert means df using .to_dict('records')
# Return converted dict
文档对过滤有这样的说法（强调我的）：
By default, these transformations are done clientside. Your Dash callbacks can respond to these modifications by listening to the data property as an Input.
您可以在此处阅读更多内容：https://dash.plotly.com/datatable/interactivity
编辑
所以经过一些实验，我意识到当 filter_action 设置为 'native' 时，DataTable 的过滤实际上并没有更新它的 data 属性。因为该属性实际上并未更新，所以使用 data 属性作为 Input 的任何回调都不会在过滤时触发。为了解决这个问题并使用 data 属性作为我们回调的输入，我们需要将 filter_action 设置为 'custom'，然后我们自己实现过滤回调。幸运的是，Dash 文档在这里有一个如何执行此操作的示例：https://dash.plotly.com/datatable/filtering
使用这些文档中的信息，我能够得到一个可以工作的示例。我将留给您来弄清楚过滤方法的内部工作原理，但以下代码在本地为我运行，并显示了在过滤主 DataTable 时 DataTable 更新的方式。我所做的唯一其他修改是对 update_means 函数，我将其从返回 list 更改为返回 dict，这是data 属性。
import dash
from dash.dependencies import Input, Output
import dash_table
import dash_core_components as dcc
import dash_html_components as html
import pandas as pd
import json
import numpy as np
from dash import Dash
# Instantiate df and calculate initial means
wide_data = pd.DataFrame(np.array([[1.24, 2.34, 3.234], [4.24, .45, .06], [7, 8, 9]]),
columns=['a', 'b', 'c'])
df = pd.DataFrame(wide_data)
df_floats = df.select_dtypes(include=['float64'])
df_floats_means = df_floats.mean(numeric_only=True)
df_floats_means = pd.DataFrame(df_floats_means)
df_floats_means_T = df_floats_means.T
# Operators used by custom filtering
operators = [['ge ', '>='],
['le ', '<='],
['lt ', '<'],
['gt ', '>'],
['ne ', '!='],
['eq ', '='],
['contains '],
['datestartswith ']]
# Initialize app
app = Dash(__name__)
# Init layout
app.layout = html.Div([
html.H1('Customer Complaints Dashboard'),
html.Label('Read and Writing Queries Filtering'),
# html.Label2('If "Read filter_query" is chosen, use operators (le, ge, <, <=, >, >=) + tag value to filter'),
# html.Label3('If "Write filter_query" is chosen, ___'),
dcc.RadioItems(
id='filter-query-read-write',
options=[
{'label': 'Read filter_query', 'value': 'read'},
],
value='read'
),
html.Br(),
# this is for write functionality
dcc.Input(id='filter-query-input', placeholder='Enter filter query'),
html.Div(id='filter-query-output'), # this is associated with read screen
html.Hr(),
dash_table.DataTable(
id='datatable-advanced-filtering',
columns=[
{'name': i, 'id': i, 'deletable': True} for i in df.columns
# omit the id column
if i != 'id'
],
data=df.to_dict('records'),
editable=True,
page_action='native',
page_size=10,
filter_action="custom",
fixed_columns={'headers': True, 'data': 1},
style_table={'minWidth': '100%'}
),
html.Hr(),
html.Div(id='datatable-query-structure', style={'whitespace': 'pre'}),
html.Hr(),
html.Label('Dataframe Means'),
dash_table.DataTable(
id='datatable-advanced-filtering2',
columns=[
{'name': i, 'id': i, 'deletable': True} for i in df_floats_means_T.columns
# omit the id column
if i != 'id'
],
data=df_floats_means_T.to_dict('records'),
editable=True,
page_action='native',
page_size=10,
filter_action="native",
fixed_columns={'headers': True, 'data': 1},
style_table={'minWidth': '100%'}
),
html.Hr(),
html.Div(id='datatable-query-structure2', style={'whitespace': 'pre'}),
html.Hr(),
])
@app.callback(
[Output('filter-query-input', 'style'),
Output('filter-query-output', 'style')],
[Input('filter-query-read-write', 'value')]
)
def query_input_output(val):
input_style = {'width': '100%'}
output_style = {}
if val == 'read':
input_style.update(display='none')
output_style.update(display='inline-block')
else:
input_style.update(display='inline-block')
output_style.update(display='none')
return input_style, output_style
@app.callback(
Output('filter-query-output', 'children'),
[Input('datatable-advanced-filtering', 'filter_query')]
)
def read_query(query):
if query is None:
return "No filter query"
return dcc.Markdown('`filter_query = "{}"`'.format(query))
# Callback to re-calculate means after filtering is
@app.callback(
Output('datatable-advanced-filtering2', 'data'),
[Input('datatable-advanced-filtering', 'data')]
)
def update_means(data):
# Calculate means from data currently stored in top datatable
dff = pd.DataFrame.from_dict(data)
df_floats = dff.select_dtypes(include=['float64'])
df_floats_means = df_floats.mean(numeric_only=True)
df_floats_means = pd.DataFrame(df_floats_means)
df_floats_means_T = df_floats_means.T
# Return means to means datatable
#THIS NOW RETURNS DICT INSTEAD OF LIST
converted_means = df_floats_means_T.to_dict('records')
return converted_means
def split_filter_part(filter_part):
'''Helper function for custom filtering'''
for operator_type in operators:
for operator in operator_type:
if operator in filter_part:
name_part, value_part = filter_part.split(operator, 1)
name = name_part[name_part.find('{') + 1: name_part.rfind('}')]
value_part = value_part.strip()
v0 = value_part[0]
if (v0 == value_part[-1] and v0 in ("'", '"', '`')):
value = value_part[1: -1].replace('\' + v0, v0)
else:
try:
value = float(value_part)
except ValueError:
value = value_part
# word operators need spaces after them in the filter string,
# but we don't want these later
return name, operator_type[0].strip(), value
return [None] * 3
@app.callback(
Output('datatable-advanced-filtering', "data"),
[Input('datatable-advanced-filtering', "filter_query")])
def update_table(filter):
'''Callback that handles custom filtering of top datatable'''
if filter is None:
return df.to_dict('records')
filtering_expressions = filter.split(' && ')
dff = df
for filter_part in filtering_expressions:
col_name, operator, filter_value = split_filter_part(filter_part)
if operator in ('eq', 'ne', 'lt', 'le', 'gt', 'ge'):
# these operators match pandas series operator method names
dff = dff.loc[getattr(dff[col_name], operator)(filter_value)]
elif operator == 'contains':
dff = dff.loc[dff[col_name].str.contains(filter_value)]
elif operator == 'datestartswith':
# this is a simplification of the front-end filtering logic,
# only works with complete fields in standard format
dff = dff.loc[dff[col_name].str.startswith(filter_value)]
return dff.to_dict('records')
if __name__ == '__main__':
app.run_server(threaded=True)
@ncascale 非常感谢，但这里有一个小问题：
def update_table(filter):
'''Callback that handles custom filtering of top datatable'''
filtering_expressions = filter.split(' && ')
dff = df
for filter_part in filtering_expressions:
col_name, operator, filter_value = split_filter_part(filter_part)
if operator in ('eq', 'ne', 'lt', 'le', 'gt', 'ge'):
# these operators match pandas series operator method names
dff = dff.loc[getattr(dff[col_name], operator)(filter_value)]
elif operator == 'contains':
dff = dff.loc[dff[col_name].str.contains(filter_value)]
elif operator == 'datestartswith':
# this is a simplification of the front-end filtering logic,
# only works with complete fields in standard format
dff = dff.loc[dff[col_name].str.startswith(filter_value)]
return dff.to_dict('records')
有效，但弹出一个错误：
AttributeError: 'NoneType' object has no attribute 'split'
这是因为我必须首先使用 If 语句检查此处传递的参数 "filter" 是否必须 None？
我正在使用 Python Dash 构建一个应用程序，我可以在其中显示破折号数据表，并将数据表（基础 pandas 数据框）的列平均值报告为破折号数据表在它下面..到目前为止，我已经尝试过：
import dash
from dash.dependencies import Input, Output
import dash_table
import dash_core_components as dcc
import dash_html_components as html
import pandas as pd
import json
wide_data = pd.DataFrame(np.array([[1.24, 2.34, 3.234], [4.24, .45, .06], [7, 8, 9]]),
columns=['a', 'b', 'c'])
df = pd.DataFrame(wide_data)
df = pd.DataFrame(wide_data)
df_floats = df.select_dtypes(include=['float64'])
df_floats_means = df_floats.mean(numeric_only=True)
df_floats_means = pd.DataFrame(df_floats_means)
df_floats_means_T = df_floats_means.T
app = dash.Dash(__name__)
app.layout = html.Div([
html.H1('Customer Complaints Dashboard'),
html.Label('Read and Writing Queries Filtering'),
# html.Label2('If "Read filter_query" is chosen, use operators (le, ge, <, <=, >, >=) + tag value to filter'),
# html.Label3('If "Write filter_query" is chosen, ___'),
dcc.RadioItems(
id='filter-query-read-write',
options=[
{'label': 'Read filter_query', 'value': 'read'},
],
value='read'
),
html.Br(),
dcc.Input(id='filter-query-input', placeholder='Enter filter query'), #this is for write functionality
html.Div(id='filter-query-output'), #this is associated with read screen
html.Hr(),
dash_table.DataTable(
id='datatable-advanced-filtering',
columns=[
{'name': i, 'id': i, 'deletable': True} for i in df.columns
# omit the id column
if i != 'id'
],
data=df.to_dict('records'),
editable=True,
page_action='native',
page_size=10,
filter_action="native",
fixed_columns={'headers': True, 'data': 1},
style_table={'minWidth': '100%'}
),
html.Hr(),
html.Div(id='datatable-query-structure', style={'whitespace': 'pre'}),
html.Hr(),
html.Label('Dataframe Means'),
dash_table.DataTable(
id='datatable-advanced-filtering2',
columns=[
{'name': i, 'id': i, 'deletable': True} for i in df_floats_means_T.columns
# omit the id column
if i != 'id'
],
data=df_floats_means_T.to_dict('records'),
editable=True,
page_action='native',
page_size=10,
# filter_action="native",
fixed_columns={'headers': True, 'data': 1},
style_table={'minWidth': '100%'}
),
html.Hr(),
html.Div(id='datatable-query-structure2', style={'whitespace': 'pre'}),
html.Hr(),
])
@app.callback(
[Output('filter-query-input', 'style'),
Output('filter-query-output', 'style')],
[Input('filter-query-read-write', 'value')]
)
def query_input_output(val):
input_style = {'width': '100%'}
output_style = {}
if val == 'read':
input_style.update(display='none')
output_style.update(display='inline-block')
else:
input_style.update(display='inline-block')
output_style.update(display='none')
return input_style, output_style
@app.callback(
Output('filter-query-output', 'children'),
[Input('datatable-advanced-filtering', 'filter_query')]
)
def read_query(query):
if query is None:
return "No filter query"
return dcc.Markdown('`filter_query = "{}"`'.format(query))
@app.callback(
Output('datatable-advanced-filtering2', 'data'),
[Input('datatable-advanced-filtering', 'data')]
)
def update_means(data):
dff = df
df_floats = dff.select_dtypes(include=['float64'])
df_floats_means = df_floats.mean(numeric_only=True)
df_floats_means = pd.DataFrame(df_floats_means)
df_floats_means_T = df_floats_means.T
converted_means = df_floats_means_T.to_dict('records')
return([converted_means])
if __name__ == '__main__':
app.run_server(threaded=True)
但我知道这是不对的，因为我在应用程序布局的这一部分进行任何过滤之前计算均值 id='datatable-advanced-filtering2'。大家能帮我回调一下吗？
我 100% 是 dash 的新手，而且我还没有看到很多关于如何正确执行此操作的指南。任何帮助将不胜感激。谢谢。
我相信您可以使用主 DataTable 的 data 属性 作为回调函数的输入，并将输出设为 data 属性你的意思是数据表。其结构可能类似于：
@app.callback(
Output('datatable-advanced-filtering2', 'data'),
[Input('datatable-advanced-filtering', 'data')]
)
def update_means(data):
# Use data to construct dataframe
# Find means using same method used at top of post
# Convert means df using .to_dict('records')
# Return converted dict
文档对过滤有这样的说法（强调我的）：
By default, these transformations are done clientside. Your Dash callbacks can respond to these modifications by listening to the data property as an Input.
您可以在此处阅读更多内容：https://dash.plotly.com/datatable/interactivity
编辑
所以经过一些实验，我意识到当 filter_action 设置为 'native' 时，DataTable 的过滤实际上并没有更新它的 data 属性。因为该属性实际上并未更新，所以使用 data 属性作为 Input 的任何回调都不会在过滤时触发。为了解决这个问题并使用 data 属性作为我们回调的输入，我们需要将 filter_action 设置为 'custom'，然后我们自己实现过滤回调。幸运的是，Dash 文档在这里有一个如何执行此操作的示例：https://dash.plotly.com/datatable/filtering
使用这些文档中的信息，我能够得到一个可以工作的示例。我将留给您来弄清楚过滤方法的内部工作原理，但以下代码在本地为我运行，并显示了在过滤主 DataTable 时 DataTable 更新的方式。我所做的唯一其他修改是对 update_means 函数，我将其从返回 list 更改为返回 dict，这是data 属性。
import dash
from dash.dependencies import Input, Output
import dash_table
import dash_core_components as dcc
import dash_html_components as html
import pandas as pd
import json
import numpy as np
from dash import Dash
# Instantiate df and calculate initial means
wide_data = pd.DataFrame(np.array([[1.24, 2.34, 3.234], [4.24, .45, .06], [7, 8, 9]]),
columns=['a', 'b', 'c'])
df = pd.DataFrame(wide_data)
df_floats = df.select_dtypes(include=['float64'])
df_floats_means = df_floats.mean(numeric_only=True)
df_floats_means = pd.DataFrame(df_floats_means)
df_floats_means_T = df_floats_means.T
# Operators used by custom filtering
operators = [['ge ', '>='],
['le ', '<='],
['lt ', '<'],
['gt ', '>'],
['ne ', '!='],
['eq ', '='],
['contains '],
['datestartswith ']]
# Initialize app
app = Dash(__name__)
# Init layout
app.layout = html.Div([
html.H1('Customer Complaints Dashboard'),
html.Label('Read and Writing Queries Filtering'),
# html.Label2('If "Read filter_query" is chosen, use operators (le, ge, <, <=, >, >=) + tag value to filter'),
# html.Label3('If "Write filter_query" is chosen, ___'),
dcc.RadioItems(
id='filter-query-read-write',
options=[
{'label': 'Read filter_query', 'value': 'read'},
],
value='read'
),
html.Br(),
# this is for write functionality
dcc.Input(id='filter-query-input', placeholder='Enter filter query'),
html.Div(id='filter-query-output'), # this is associated with read screen
html.Hr(),
dash_table.DataTable(
id='datatable-advanced-filtering',
columns=[
{'name': i, 'id': i, 'deletable': True} for i in df.columns
# omit the id column
if i != 'id'
],
data=df.to_dict('records'),
editable=True,
page_action='native',
page_size=10,
filter_action="custom",
fixed_columns={'headers': True, 'data': 1},
style_table={'minWidth': '100%'}
),
html.Hr(),
html.Div(id='datatable-query-structure', style={'whitespace': 'pre'}),
html.Hr(),
html.Label('Dataframe Means'),
dash_table.DataTable(
id='datatable-advanced-filtering2',
columns=[
{'name': i, 'id': i, 'deletable': True} for i in df_floats_means_T.columns
# omit the id column
if i != 'id'
],
data=df_floats_means_T.to_dict('records'),
editable=True,
page_action='native',
page_size=10,
filter_action="native",
fixed_columns={'headers': True, 'data': 1},
style_table={'minWidth': '100%'}
),
html.Hr(),
html.Div(id='datatable-query-structure2', style={'whitespace': 'pre'}),
html.Hr(),
])
@app.callback(
[Output('filter-query-input', 'style'),
Output('filter-query-output', 'style')],
[Input('filter-query-read-write', 'value')]
)
def query_input_output(val):
input_style = {'width': '100%'}
output_style = {}
if val == 'read':
input_style.update(display='none')
output_style.update(display='inline-block')
else:
input_style.update(display='inline-block')
output_style.update(display='none')
return input_style, output_style
@app.callback(
Output('filter-query-output', 'children'),
[Input('datatable-advanced-filtering', 'filter_query')]
)
def read_query(query):
if query is None:
return "No filter query"
return dcc.Markdown('`filter_query = "{}"`'.format(query))
# Callback to re-calculate means after filtering is
@app.callback(
Output('datatable-advanced-filtering2', 'data'),
[Input('datatable-advanced-filtering', 'data')]
)
def update_means(data):
# Calculate means from data currently stored in top datatable
dff = pd.DataFrame.from_dict(data)
df_floats = dff.select_dtypes(include=['float64'])
df_floats_means = df_floats.mean(numeric_only=True)
df_floats_means = pd.DataFrame(df_floats_means)
df_floats_means_T = df_floats_means.T
# Return means to means datatable
#THIS NOW RETURNS DICT INSTEAD OF LIST
converted_means = df_floats_means_T.to_dict('records')
return converted_means
def split_filter_part(filter_part):
'''Helper function for custom filtering'''
for operator_type in operators:
for operator in operator_type:
if operator in filter_part:
name_part, value_part = filter_part.split(operator, 1)
name = name_part[name_part.find('{') + 1: name_part.rfind('}')]
value_part = value_part.strip()
v0 = value_part[0]
if (v0 == value_part[-1] and v0 in ("'", '"', '`')):
value = value_part[1: -1].replace('\' + v0, v0)
else:
try:
value = float(value_part)
except ValueError:
value = value_part
# word operators need spaces after them in the filter string,
# but we don't want these later
return name, operator_type[0].strip(), value
return [None] * 3
@app.callback(
Output('datatable-advanced-filtering', "data"),
[Input('datatable-advanced-filtering', "filter_query")])
def update_table(filter):
'''Callback that handles custom filtering of top datatable'''
if filter is None:
return df.to_dict('records')
filtering_expressions = filter.split(' && ')
dff = df
for filter_part in filtering_expressions:
col_name, operator, filter_value = split_filter_part(filter_part)
if operator in ('eq', 'ne', 'lt', 'le', 'gt', 'ge'):
# these operators match pandas series operator method names
dff = dff.loc[getattr(dff[col_name], operator)(filter_value)]
elif operator == 'contains':
dff = dff.loc[dff[col_name].str.contains(filter_value)]
elif operator == 'datestartswith':
# this is a simplification of the front-end filtering logic,
# only works with complete fields in standard format
dff = dff.loc[dff[col_name].str.startswith(filter_value)]
return dff.to_dict('records')
if __name__ == '__main__':
app.run_server(threaded=True)
@ncascale 非常感谢，但这里有一个小问题：
def update_table(filter):
'''Callback that handles custom filtering of top datatable'''
filtering_expressions = filter.split(' && ')
dff = df
for filter_part in filtering_expressions:
col_name, operator, filter_value = split_filter_part(filter_part)
if operator in ('eq', 'ne', 'lt', 'le', 'gt', 'ge'):
# these operators match pandas series operator method names
dff = dff.loc[getattr(dff[col_name], operator)(filter_value)]
elif operator == 'contains':
dff = dff.loc[dff[col_name].str.contains(filter_value)]
elif operator == 'datestartswith':
# this is a simplification of the front-end filtering logic,
# only works with complete fields in standard format
dff = dff.loc[dff[col_name].str.startswith(filter_value)]
return dff.to_dict('records')
有效，但弹出一个错误：
AttributeError: 'NoneType' object has no attribute 'split'
这是因为我必须首先使用 If 语句检查此处传递的参数 "filter" 是否必须 None？