# -*- coding: utf-8 -*-
import numpy as np
from .. import metaarray as metaarray
from ..Qt import QtCore
from .GraphicsObject import GraphicsObject
from .PlotCurveItem import PlotCurveItem
from .ScatterPlotItem import ScatterPlotItem
from .. import functions as fn
from .. import debug as debug
from .. import getConfigOption
[docs]class PlotDataItem(GraphicsObject):
"""
**Bases:** :class:`GraphicsObject <pyqtgraph.GraphicsObject>`
GraphicsItem for displaying plot curves, scatter plots, or both.
While it is possible to use :class:`PlotCurveItem <pyqtgraph.PlotCurveItem>` or
:class:`ScatterPlotItem <pyqtgraph.ScatterPlotItem>` individually, this class
provides a unified interface to both. Instances of :class:`PlotDataItem` are
usually created by plot() methods such as :func:`pyqtgraph.plot` and
:func:`PlotItem.plot() <pyqtgraph.PlotItem.plot>`.
============================== ==============================================
**Signals:**
sigPlotChanged(self) Emitted when the data in this item is updated.
sigClicked(self) Emitted when the item is clicked.
sigPointsClicked(self, points) Emitted when a plot point is clicked
Sends the list of points under the mouse.
============================== ==============================================
"""
sigPlotChanged = QtCore.Signal(object)
sigClicked = QtCore.Signal(object)
sigPointsClicked = QtCore.Signal(object, object)
[docs] def __init__(self, *args, **kargs):
"""
There are many different ways to create a PlotDataItem:
**Data initialization arguments:** (x,y data only)
=================================== ======================================
PlotDataItem(xValues, yValues) x and y values may be any sequence
(including ndarray) of real numbers
PlotDataItem(yValues) y values only -- x will be
automatically set to range(len(y))
PlotDataItem(x=xValues, y=yValues) x and y given by keyword arguments
PlotDataItem(ndarray(Nx2)) numpy array with shape (N, 2) where
``x=data[:,0]`` and ``y=data[:,1]``
=================================== ======================================
**Data initialization arguments:** (x,y data AND may include spot style)
============================ =========================================
PlotDataItem(recarray) numpy array with ``dtype=[('x', float),
('y', float), ...]``
PlotDataItem(list-of-dicts) ``[{'x': x, 'y': y, ...}, ...]``
PlotDataItem(dict-of-lists) ``{'x': [...], 'y': [...], ...}``
PlotDataItem(MetaArray) 1D array of Y values with X sepecified as
axis values OR 2D array with a column 'y'
and extra columns as needed.
============================ =========================================
**Line style keyword arguments:**
============ ==============================================================================
connect Specifies how / whether vertexes should be connected. See
:func:`arrayToQPath() <pyqtgraph.arrayToQPath>`
pen Pen to use for drawing line between points.
Default is solid grey, 1px width. Use None to disable line drawing.
May be any single argument accepted by :func:`mkPen() <pyqtgraph.mkPen>`
shadowPen Pen for secondary line to draw behind the primary line. disabled by default.
May be any single argument accepted by :func:`mkPen() <pyqtgraph.mkPen>`
fillLevel Fill the area between the curve and fillLevel
fillOutline (bool) If True, an outline surrounding the *fillLevel* area is drawn.
fillBrush Fill to use when fillLevel is specified.
May be any single argument accepted by :func:`mkBrush() <pyqtgraph.mkBrush>`
stepMode If True, two orthogonal lines are drawn for each sample
as steps. This is commonly used when drawing histograms.
Note that in this case, ``len(x) == len(y) + 1``
(added in version 0.9.9)
============ ==============================================================================
**Point style keyword arguments:** (see :func:`ScatterPlotItem.setData() <pyqtgraph.ScatterPlotItem.setData>` for more information)
============ =====================================================
symbol Symbol to use for drawing points OR list of symbols,
one per point. Default is no symbol.
Options are o, s, t, d, +, or any QPainterPath
symbolPen Outline pen for drawing points OR list of pens, one
per point. May be any single argument accepted by
:func:`mkPen() <pyqtgraph.mkPen>`
symbolBrush Brush for filling points OR list of brushes, one per
point. May be any single argument accepted by
:func:`mkBrush() <pyqtgraph.mkBrush>`
symbolSize Diameter of symbols OR list of diameters.
pxMode (bool) If True, then symbolSize is specified in
pixels. If False, then symbolSize is
specified in data coordinates.
============ =====================================================
**Optimization keyword arguments:**
================ =====================================================================
antialias (bool) By default, antialiasing is disabled to improve performance.
Note that in some cases (in particluar, when pxMode=True), points
will be rendered antialiased even if this is set to False.
decimate deprecated.
downsample (int) Reduce the number of samples displayed by this value
downsampleMethod 'subsample': Downsample by taking the first of N samples.
This method is fastest and least accurate.
'mean': Downsample by taking the mean of N samples.
'peak': Downsample by drawing a saw wave that follows the min
and max of the original data. This method produces the best
visual representation of the data but is slower.
autoDownsample (bool) If True, resample the data before plotting to avoid plotting
multiple line segments per pixel. This can improve performance when
viewing very high-density data, but increases the initial overhead
and memory usage.
clipToView (bool) If True, only plot data that is visible within the X range of
the containing ViewBox. This can improve performance when plotting
very large data sets where only a fraction of the data is visible
at any time.
identical *deprecated*
================ =====================================================================
**Meta-info keyword arguments:**
========== ================================================
name name of dataset. This would appear in a legend
========== ================================================
"""
GraphicsObject.__init__(self)
self.setFlag(self.ItemHasNoContents)
self.xData = None
self.yData = None
self.xDisp = None
self.yDisp = None
#self.dataMask = None
#self.curves = []
#self.scatters = []
self.curve = PlotCurveItem()
self.scatter = ScatterPlotItem()
self.curve.setParentItem(self)
self.scatter.setParentItem(self)
self.curve.sigClicked.connect(self.curveClicked)
self.scatter.sigClicked.connect(self.scatterClicked)
#self.clear()
self.opts = {
'connect': 'all',
'fftMode': False,
'logMode': [False, False],
'alphaHint': 1.0,
'alphaMode': False,
'pen': (200,200,200),
'shadowPen': None,
'fillLevel': None,
'fillOutline': False,
'fillBrush': None,
'stepMode': None,
'symbol': None,
'symbolSize': 10,
'symbolPen': (200,200,200),
'symbolBrush': (50, 50, 150),
'pxMode': True,
'antialias': getConfigOption('antialias'),
'pointMode': None,
'downsample': 1,
'autoDownsample': False,
'downsampleMethod': 'peak',
'autoDownsampleFactor': 5., # draw ~5 samples per pixel
'clipToView': False,
'data': None,
}
self.setData(*args, **kargs)
def implements(self, interface=None):
ints = ['plotData']
if interface is None:
return ints
return interface in ints
def name(self):
return self.opts.get('name', None)
def boundingRect(self):
return QtCore.QRectF() ## let child items handle this
def setAlpha(self, alpha, auto):
if self.opts['alphaHint'] == alpha and self.opts['alphaMode'] == auto:
return
self.opts['alphaHint'] = alpha
self.opts['alphaMode'] = auto
self.setOpacity(alpha)
#self.update()
def setFftMode(self, mode):
if self.opts['fftMode'] == mode:
return
self.opts['fftMode'] = mode
self.xDisp = self.yDisp = None
self.xClean = self.yClean = None
self.updateItems()
self.informViewBoundsChanged()
def setLogMode(self, xMode, yMode):
if self.opts['logMode'] == [xMode, yMode]:
return
self.opts['logMode'] = [xMode, yMode]
self.xDisp = self.yDisp = None
self.xClean = self.yClean = None
self.updateItems()
self.informViewBoundsChanged()
def setPointMode(self, mode):
if self.opts['pointMode'] == mode:
return
self.opts['pointMode'] = mode
self.update()
[docs] def setPen(self, *args, **kargs):
"""
| Sets the pen used to draw lines between points.
| *pen* can be a QPen or any argument accepted by :func:`pyqtgraph.mkPen() <pyqtgraph.mkPen>`
"""
pen = fn.mkPen(*args, **kargs)
self.opts['pen'] = pen
#self.curve.setPen(pen)
#for c in self.curves:
#c.setPen(pen)
#self.update()
self.updateItems()
[docs] def setShadowPen(self, *args, **kargs):
"""
| Sets the shadow pen used to draw lines between points (this is for enhancing contrast or
emphacizing data).
| This line is drawn behind the primary pen (see :func:`setPen() <pyqtgraph.PlotDataItem.setPen>`)
and should generally be assigned greater width than the primary pen.
| *pen* can be a QPen or any argument accepted by :func:`pyqtgraph.mkPen() <pyqtgraph.mkPen>`
"""
pen = fn.mkPen(*args, **kargs)
self.opts['shadowPen'] = pen
#for c in self.curves:
#c.setPen(pen)
#self.update()
self.updateItems()
def setFillBrush(self, *args, **kargs):
brush = fn.mkBrush(*args, **kargs)
if self.opts['fillBrush'] == brush:
return
self.opts['fillBrush'] = brush
self.updateItems()
def setBrush(self, *args, **kargs):
return self.setFillBrush(*args, **kargs)
def setFillLevel(self, level):
if self.opts['fillLevel'] == level:
return
self.opts['fillLevel'] = level
self.updateItems()
def setSymbol(self, symbol):
if self.opts['symbol'] == symbol:
return
self.opts['symbol'] = symbol
#self.scatter.setSymbol(symbol)
self.updateItems()
def setSymbolPen(self, *args, **kargs):
pen = fn.mkPen(*args, **kargs)
if self.opts['symbolPen'] == pen:
return
self.opts['symbolPen'] = pen
#self.scatter.setSymbolPen(pen)
self.updateItems()
def setSymbolBrush(self, *args, **kargs):
brush = fn.mkBrush(*args, **kargs)
if self.opts['symbolBrush'] == brush:
return
self.opts['symbolBrush'] = brush
#self.scatter.setSymbolBrush(brush)
self.updateItems()
def setSymbolSize(self, size):
if self.opts['symbolSize'] == size:
return
self.opts['symbolSize'] = size
#self.scatter.setSymbolSize(symbolSize)
self.updateItems()
[docs] def setDownsampling(self, ds=None, auto=None, method=None):
"""
Set the downsampling mode of this item. Downsampling reduces the number
of samples drawn to increase performance.
============== =================================================================
**Arguments:**
ds (int) Reduce visible plot samples by this factor. To disable,
set ds=1.
auto (bool) If True, automatically pick *ds* based on visible range
mode 'subsample': Downsample by taking the first of N samples.
This method is fastest and least accurate.
'mean': Downsample by taking the mean of N samples.
'peak': Downsample by drawing a saw wave that follows the min
and max of the original data. This method produces the best
visual representation of the data but is slower.
============== =================================================================
"""
changed = False
if ds is not None:
if self.opts['downsample'] != ds:
changed = True
self.opts['downsample'] = ds
if auto is not None and self.opts['autoDownsample'] != auto:
self.opts['autoDownsample'] = auto
changed = True
if method is not None:
if self.opts['downsampleMethod'] != method:
changed = True
self.opts['downsampleMethod'] = method
if changed:
self.xDisp = self.yDisp = None
self.updateItems()
def setClipToView(self, clip):
if self.opts['clipToView'] == clip:
return
self.opts['clipToView'] = clip
self.xDisp = self.yDisp = None
self.updateItems()
[docs] def setData(self, *args, **kargs):
"""
Clear any data displayed by this item and display new data.
See :func:`__init__() <pyqtgraph.PlotDataItem.__init__>` for details; it accepts the same arguments.
"""
#self.clear()
profiler = debug.Profiler()
y = None
x = None
if len(args) == 1:
data = args[0]
dt = dataType(data)
if dt == 'empty':
pass
elif dt == 'listOfValues':
y = np.array(data)
elif dt == 'Nx2array':
x = data[:,0]
y = data[:,1]
elif dt == 'recarray' or dt == 'dictOfLists':
if 'x' in data:
x = np.array(data['x'])
if 'y' in data:
y = np.array(data['y'])
elif dt == 'listOfDicts':
if 'x' in data[0]:
x = np.array([d.get('x',None) for d in data])
if 'y' in data[0]:
y = np.array([d.get('y',None) for d in data])
for k in ['data', 'symbolSize', 'symbolPen', 'symbolBrush', 'symbolShape']:
if k in data:
kargs[k] = [d.get(k, None) for d in data]
elif dt == 'MetaArray':
y = data.view(np.ndarray)
x = data.xvals(0).view(np.ndarray)
else:
raise Exception('Invalid data type %s' % type(data))
elif len(args) == 2:
seq = ('listOfValues', 'MetaArray', 'empty')
dtyp = dataType(args[0]), dataType(args[1])
if dtyp[0] not in seq or dtyp[1] not in seq:
raise Exception('When passing two unnamed arguments, both must be a list or array of values. (got %s, %s)' % (str(type(args[0])), str(type(args[1]))))
if not isinstance(args[0], np.ndarray):
#x = np.array(args[0])
if dtyp[0] == 'MetaArray':
x = args[0].asarray()
else:
x = np.array(args[0])
else:
x = args[0].view(np.ndarray)
if not isinstance(args[1], np.ndarray):
#y = np.array(args[1])
if dtyp[1] == 'MetaArray':
y = args[1].asarray()
else:
y = np.array(args[1])
else:
y = args[1].view(np.ndarray)
if 'x' in kargs:
x = kargs['x']
if 'y' in kargs:
y = kargs['y']
profiler('interpret data')
## pull in all style arguments.
## Use self.opts to fill in anything not present in kargs.
if 'name' in kargs:
self.opts['name'] = kargs['name']
if 'connect' in kargs:
self.opts['connect'] = kargs['connect']
## if symbol pen/brush are given with no symbol, then assume symbol is 'o'
if 'symbol' not in kargs and ('symbolPen' in kargs or 'symbolBrush' in kargs or 'symbolSize' in kargs):
kargs['symbol'] = 'o'
if 'brush' in kargs:
kargs['fillBrush'] = kargs['brush']
for k in list(self.opts.keys()):
if k in kargs:
self.opts[k] = kargs[k]
#curveArgs = {}
#for k in ['pen', 'shadowPen', 'fillLevel', 'brush']:
#if k in kargs:
#self.opts[k] = kargs[k]
#curveArgs[k] = self.opts[k]
#scatterArgs = {}
#for k,v in [('symbolPen','pen'), ('symbolBrush','brush'), ('symbol','symbol')]:
#if k in kargs:
#self.opts[k] = kargs[k]
#scatterArgs[v] = self.opts[k]
if y is None:
self.updateItems()
profiler('update items')
return
if y is not None and x is None:
x = np.arange(len(y))
if not isinstance(x, np.ndarray):
x = np.array(x)
if not isinstance(y, np.ndarray):
y = np.array(y)
self.xData = x.view(np.ndarray) ## one last check to make sure there are no MetaArrays getting by
self.yData = y.view(np.ndarray)
self.xClean = self.yClean = None
self.xDisp = None
self.yDisp = None
profiler('set data')
self.updateItems()
profiler('update items')
self.informViewBoundsChanged()
#view = self.getViewBox()
#if view is not None:
#view.itemBoundsChanged(self) ## inform view so it can update its range if it wants
self.sigPlotChanged.emit(self)
profiler('emit')
def updateItems(self):
curveArgs = {}
for k,v in [('pen','pen'), ('shadowPen','shadowPen'), ('fillLevel','fillLevel'), ('fillOutline', 'fillOutline'), ('fillBrush', 'brush'), ('antialias', 'antialias'), ('connect', 'connect'), ('stepMode', 'stepMode')]:
curveArgs[v] = self.opts[k]
scatterArgs = {}
for k,v in [('symbolPen','pen'), ('symbolBrush','brush'), ('symbol','symbol'), ('symbolSize', 'size'), ('data', 'data'), ('pxMode', 'pxMode'), ('antialias', 'antialias')]:
if k in self.opts:
scatterArgs[v] = self.opts[k]
x,y = self.getData()
#scatterArgs['mask'] = self.dataMask
if curveArgs['pen'] is not None or (curveArgs['brush'] is not None and curveArgs['fillLevel'] is not None):
self.curve.setData(x=x, y=y, **curveArgs)
self.curve.show()
else:
self.curve.hide()
if scatterArgs['symbol'] is not None:
if self.opts.get('stepMode', False) is True:
x = 0.5 * (x[:-1] + x[1:])
self.scatter.setData(x=x, y=y, **scatterArgs)
self.scatter.show()
else:
self.scatter.hide()
def getData(self):
if self.xData is None:
return (None, None)
if self.xDisp is None:
x = self.xData
y = self.yData
if self.opts['fftMode']:
x,y = self._fourierTransform(x, y)
# Ignore the first bin for fft data if we have a logx scale
if self.opts['logMode'][0]:
x=x[1:]
y=y[1:]
with np.errstate(divide='ignore'):
if self.opts['logMode'][0]:
x = np.log10(x)
if self.opts['logMode'][1]:
y = np.log10(y)
ds = self.opts['downsample']
if not isinstance(ds, int):
ds = 1
if self.opts['autoDownsample']:
# this option presumes that x-values have uniform spacing
range = self.viewRect()
if range is not None and len(x) > 1:
dx = float(x[-1]-x[0]) / (len(x)-1)
if dx != 0.0:
x0 = (range.left()-x[0]) / dx
x1 = (range.right()-x[0]) / dx
width = self.getViewBox().width()
if width != 0.0:
ds = int(max(1, int((x1-x0) / (width*self.opts['autoDownsampleFactor']))))
## downsampling is expensive; delay until after clipping.
if self.opts['clipToView']:
view = self.getViewBox()
if view is None or not view.autoRangeEnabled()[0]:
# this option presumes that x-values are in increasing order
range = self.viewRect()
if range is not None and len(x) > 1:
# clip to visible region extended by downsampling value, assuming
# uniform spacing of x-values, has O(1) performance
dx = float(x[-1]-x[0]) / (len(x)-1)
x0 = np.clip(int((range.left()-x[0])/dx) - 1*ds, 0, len(x)-1)
x1 = np.clip(int((range.right()-x[0])/dx) + 2*ds, 0, len(x)-1)
# if data has been clipped too strongly (in case of non-uniform
# spacing of x-values), refine the clipping region as required
# worst case performance: O(log(n))
# best case performance: O(1)
if x[x0] > range.left():
x0 = np.searchsorted(x, range.left()) - 1*ds
x0 = np.clip(x0, a_min=0, a_max=len(x))
if x[x1] < range.right():
x1 = np.searchsorted(x, range.right()) + 2*ds
x1 = np.clip(x1, a_min=0, a_max=len(x))
x = x[x0:x1]
y = y[x0:x1]
if ds > 1:
if self.opts['downsampleMethod'] == 'subsample':
x = x[::ds]
y = y[::ds]
elif self.opts['downsampleMethod'] == 'mean':
n = len(x) // ds
x = x[:n*ds:ds]
y = y[:n*ds].reshape(n,ds).mean(axis=1)
elif self.opts['downsampleMethod'] == 'peak':
n = len(x) // ds
x1 = np.empty((n,2))
x1[:] = x[:n*ds:ds,np.newaxis]
x = x1.reshape(n*2)
y1 = np.empty((n,2))
y2 = y[:n*ds].reshape((n, ds))
y1[:,0] = y2.max(axis=1)
y1[:,1] = y2.min(axis=1)
y = y1.reshape(n*2)
self.xDisp = x
self.yDisp = y
return self.xDisp, self.yDisp
[docs] def dataBounds(self, ax, frac=1.0, orthoRange=None):
"""
Returns the range occupied by the data (along a specific axis) in this item.
This method is called by ViewBox when auto-scaling.
=============== =============================================================
**Arguments:**
ax (0 or 1) the axis for which to return this item's data range
frac (float 0.0-1.0) Specifies what fraction of the total data
range to return. By default, the entire range is returned.
This allows the ViewBox to ignore large spikes in the data
when auto-scaling.
orthoRange ([min,max] or None) Specifies that only the data within the
given range (orthogonal to *ax*) should me measured when
returning the data range. (For example, a ViewBox might ask
what is the y-range of all data with x-values between min
and max)
=============== =============================================================
"""
range = [None, None]
if self.curve.isVisible():
range = self.curve.dataBounds(ax, frac, orthoRange)
elif self.scatter.isVisible():
r2 = self.scatter.dataBounds(ax, frac, orthoRange)
range = [
r2[0] if range[0] is None else (range[0] if r2[0] is None else min(r2[0], range[0])),
r2[1] if range[1] is None else (range[1] if r2[1] is None else min(r2[1], range[1]))
]
return range
[docs] def pixelPadding(self):
"""
Return the size in pixels that this item may draw beyond the values returned by dataBounds().
This method is called by ViewBox when auto-scaling.
"""
pad = 0
if self.curve.isVisible():
pad = max(pad, self.curve.pixelPadding())
elif self.scatter.isVisible():
pad = max(pad, self.scatter.pixelPadding())
return pad
def clear(self):
#for i in self.curves+self.scatters:
#if i.scene() is not None:
#i.scene().removeItem(i)
#self.curves = []
#self.scatters = []
self.xData = None
self.yData = None
#self.xClean = None
#self.yClean = None
self.xDisp = None
self.yDisp = None
self.curve.clear()
self.scatter.clear()
def appendData(self, *args, **kargs):
pass
def curveClicked(self):
self.sigClicked.emit(self)
def scatterClicked(self, plt, points):
self.sigClicked.emit(self)
self.sigPointsClicked.emit(self, points)
def viewRangeChanged(self):
# view range has changed; re-plot if needed
if self.opts['clipToView'] or self.opts['autoDownsample']:
self.xDisp = self.yDisp = None
self.updateItems()
def _fourierTransform(self, x, y):
## Perform fourier transform. If x values are not sampled uniformly,
## then use np.interp to resample before taking fft.
dx = np.diff(x)
uniform = not np.any(np.abs(dx-dx[0]) > (abs(dx[0]) / 1000.))
if not uniform:
x2 = np.linspace(x[0], x[-1], len(x))
y = np.interp(x2, x, y)
x = x2
n = y.size
f = np.fft.rfft(y) / n
d = float(x[-1]-x[0]) / (len(x)-1)
x = np.fft.rfftfreq(n, d)
y = np.abs(f)
return x, y
def dataType(obj):
if hasattr(obj, '__len__') and len(obj) == 0:
return 'empty'
if isinstance(obj, dict):
return 'dictOfLists'
elif isSequence(obj):
first = obj[0]
if (hasattr(obj, 'implements') and obj.implements('MetaArray')):
return 'MetaArray'
elif isinstance(obj, np.ndarray):
if obj.ndim == 1:
if obj.dtype.names is None:
return 'listOfValues'
else:
return 'recarray'
elif obj.ndim == 2 and obj.dtype.names is None and obj.shape[1] == 2:
return 'Nx2array'
else:
raise Exception('array shape must be (N,) or (N,2); got %s instead' % str(obj.shape))
elif isinstance(first, dict):
return 'listOfDicts'
else:
return 'listOfValues'
def isSequence(obj):
return hasattr(obj, '__iter__') or isinstance(obj, np.ndarray) or (hasattr(obj, 'implements') and obj.implements('MetaArray'))
#class TableData:
#"""
#Class for presenting multiple forms of tabular data through a consistent interface.
#May contain:
#- numpy record array
#- list-of-dicts (all dicts are _not_ required to have the same keys)
#- dict-of-lists
#- dict (single record)
#Note: if all the values in this record are lists, it will be interpreted as multiple records
#Data can be accessed and modified by column, by row, or by value
#data[columnName]
#data[rowId]
#data[columnName, rowId] = value
#data[columnName] = [value, value, ...]
#data[rowId] = {columnName: value, ...}
#"""
#def __init__(self, data):
#self.data = data
#if isinstance(data, np.ndarray):
#self.mode = 'array'
#elif isinstance(data, list):
#self.mode = 'list'
#elif isinstance(data, dict):
#types = set(map(type, data.values()))
### dict may be a dict-of-lists or a single record
#types -= set([list, np.ndarray]) ## if dict contains any non-sequence values, it is probably a single record.
#if len(types) != 0:
#self.data = [self.data]
#self.mode = 'list'
#else:
#self.mode = 'dict'
#elif isinstance(data, TableData):
#self.data = data.data
#self.mode = data.mode
#else:
#raise TypeError(type(data))
#for fn in ['__getitem__', '__setitem__']:
#setattr(self, fn, getattr(self, '_TableData'+fn+self.mode))
#def originalData(self):
#return self.data
#def toArray(self):
#if self.mode == 'array':
#return self.data
#if len(self) < 1:
##return np.array([]) ## need to return empty array *with correct columns*, but this is very difficult, so just return None
#return None
#rec1 = self[0]
#dtype = functions.suggestRecordDType(rec1)
##print rec1, dtype
#arr = np.empty(len(self), dtype=dtype)
#arr[0] = tuple(rec1.values())
#for i in xrange(1, len(self)):
#arr[i] = tuple(self[i].values())
#return arr
#def __getitem__array(self, arg):
#if isinstance(arg, tuple):
#return self.data[arg[0]][arg[1]]
#else:
#return self.data[arg]
#def __getitem__list(self, arg):
#if isinstance(arg, basestring):
#return [d.get(arg, None) for d in self.data]
#elif isinstance(arg, int):
#return self.data[arg]
#elif isinstance(arg, tuple):
#arg = self._orderArgs(arg)
#return self.data[arg[0]][arg[1]]
#else:
#raise TypeError(type(arg))
#def __getitem__dict(self, arg):
#if isinstance(arg, basestring):
#return self.data[arg]
#elif isinstance(arg, int):
#return dict([(k, v[arg]) for k, v in self.data.items()])
#elif isinstance(arg, tuple):
#arg = self._orderArgs(arg)
#return self.data[arg[1]][arg[0]]
#else:
#raise TypeError(type(arg))
#def __setitem__array(self, arg, val):
#if isinstance(arg, tuple):
#self.data[arg[0]][arg[1]] = val
#else:
#self.data[arg] = val
#def __setitem__list(self, arg, val):
#if isinstance(arg, basestring):
#if len(val) != len(self.data):
#raise Exception("Values (%d) and data set (%d) are not the same length." % (len(val), len(self.data)))
#for i, rec in enumerate(self.data):
#rec[arg] = val[i]
#elif isinstance(arg, int):
#self.data[arg] = val
#elif isinstance(arg, tuple):
#arg = self._orderArgs(arg)
#self.data[arg[0]][arg[1]] = val
#else:
#raise TypeError(type(arg))
#def __setitem__dict(self, arg, val):
#if isinstance(arg, basestring):
#if len(val) != len(self.data[arg]):
#raise Exception("Values (%d) and data set (%d) are not the same length." % (len(val), len(self.data[arg])))
#self.data[arg] = val
#elif isinstance(arg, int):
#for k in self.data:
#self.data[k][arg] = val[k]
#elif isinstance(arg, tuple):
#arg = self._orderArgs(arg)
#self.data[arg[1]][arg[0]] = val
#else:
#raise TypeError(type(arg))
#def _orderArgs(self, args):
### return args in (int, str) order
#if isinstance(args[0], basestring):
#return (args[1], args[0])
#else:
#return args
#def __iter__(self):
#for i in xrange(len(self)):
#yield self[i]
#def __len__(self):
#if self.mode == 'array' or self.mode == 'list':
#return len(self.data)
#else:
#return max(map(len, self.data.values()))
#def columnNames(self):
#"""returns column names in no particular order"""
#if self.mode == 'array':
#return self.data.dtype.names
#elif self.mode == 'list':
#names = set()
#for row in self.data:
#names.update(row.keys())
#return list(names)
#elif self.mode == 'dict':
#return self.data.keys()
#def keys(self):
#return self.columnNames()