Question mark in the Ipython
Get the relevant descriptive information
%run System files
Execute a file
Ipython mode command
%magic Show all the Magic commands
%hist Command History input information
%PDB automatically enters the debugger after an exception occurs
%reset Delete all variables or names in the current namespace
%WHO displays Ipython variables already defined in the current namespace
%time statemnent gives code execution time
%timeit statement executes code multiple times, calculates average execution time
Basic methods of reading data
Dimensions: the organization of data
One-dimensional data list array
A set of data can have different data types in the structure list, with the same array type
Two-bit data combined form of multiple one-dimensional data
Extended formation of multidimensional data in sexual dimension
High-dimensional data dictionary
One data list (ordered) collection (unordered)
High-dimensional data dictionary form or data representation format (JSON XML YAML)
NumPy
Np.ndarray
Import NumPy as NP
Make one-dimensional vector image variable
One dimension data type is often the same to save memory and time
Ndarray
Actual data
Metadata that describes the data
The following table starts from 0
Dimensions of Axis axis data
Number of rank rank axes
Ndarray properties of an object
. Ndim number of rank axes or number of dimensions
Scale n row m column of the. Shape object
. Size object Element number
The element type of the. Dtype Object
. itemsize the size of each element in bytes
Element type of Ndarray
bool
INTC 32 or 64
int p The integer used for the index is consistent with sszie_t in C int32 or Int64
Int8 int16 int32 Int64
Uint8-16-32-64
float16-32-64/
Complex64 Real parts are 32-bit floating-point type
complex128 Real parts are 64-bit floating-point type
When each element row is different, the non-homogeneous object avoids using it as much as possible.
Methods for generating Ndarray
1 creating from List tuples
Np.array (, Dtype=np.float32) can specify type
2 using functions to create
Np.arange () elements from 0 to n-1
Np.ones (Shape) generates a full 1 type based on shape, which is a tuple type
Np.zeros (Shape) Ibid.
Np.full (shape,val) generates the Authority Val's
Np.eye (n) n-Order unit matrix
Np.ones_like (a) similar in shape
Np.zeros_like (a)
Np.zeros_like (A,val)
Np.linspace (1,10,4, endpoint=false) starting position the Germplasm element contains several elements
Np.logspace (1,2,base=4) 4^1-4^2 geometric series
Np.concatenate ((b)) Merge two Nparray
Methods of Ndarray
. Reshape (Shape) does not change the array element, returns an array of shape shapes, the original array does not change
. Resize (Shape) modifies the original array (the original array is changed) resize does not accept negative numbers ( -1) reshape can
. swapaxes (AX1,AX2) swap two dimensions of n dimension
. Flatten () to reduce the dimension of the array, return a one-dimensional arrays, the original array unchanged
Type transformation of Ndarray
New_a=a.astype (New_type)//Np.int specifically that type has PY self-determination
(Do not change the original data) create a new array, copy
Array goto List
. ToList ()
Much slower.
Array Manipulation index slices
A[1:4:2] Start number: termination Number: Step
Does not contain a final value
a[0,1,2] This is OK.
A[-1,-2,-3] This is a right-to-left index
A[:,1,-3]
A[:,1:3,:]
A[:,:,::2] The last dimension is step 2
Operation of Ndarray
Scalar operations
1 each element in the array is calculated with it
A=a/a.mean ()
Scalar elements
Np.abs (x)
Np.fabs ()
NP.SQRT ()
Np.squar ()
Np.log () np.log10 () np.log2 ()
Np.ceil () Np.floor ()
Np.rint () rounding
NP.MODF () returns the decimal and integer numbers of the array as two separate arrays
Np.cos cosh sin sinh tan tanh
Np.exp ()
Np.sign ()
+-*/**
Np.maximum (x, y) Np.fmax ()
Np.minimum (x, y) np.fmin () to find the corresponding maximum minimum value
Np.mod (x, y) elements and modulo operations
Np.copysign (x, y) copies the symbol of the Y element to the corresponding element in the corresponding array x
><==!= generating an array of type bool
Python data analysis and presentation [first week]