The larry project is a data array that let's you label each row and column. A two-dimensional larry, for example, contains a NumPy array to hold the data and a list of row labels and column labels. A larry can have any dimension.
Alignment by label is automatic when you add (or subtract, multiply, divide) two larrys.
larry has built-in methods such as movingsum, ranking, merge, shuffle, zscore, demean, lag as well as typical Numpy methods like sum, max, std, sign, clip. NaNs are treated as missing data.
You can archive larrys in HDF5 format using save and load or using a dictionary-like interface.
larry adds the convenience of labels, provides many built-in methods, and let's you use many of your existing NumPy array functions.
docs http://larry.sourceforge.net code https://code.launchpad.net/~kwgoodman/larry/trunk mail http://groups.google.ca/group/pystatsmodels
What's New in This Release: [ read full changelog ]
New larry methods:
· astype: Copy of larry cast to specified type
· geometric_mean: new method based on existing array function
· la.util.resample.cross_validation: k-fold cross validation index iterator
· la.util.resample.bootstrap: bootstrap index iterator
· la.util.misc.listmap: O(n) version of map(list1.index, list2)
· la/src/clistmap.pyx: Cython version of listmap with python fallback
· Major performance boost in most larry methods!
· You can now use an optional dtype when creating larrys
· You can now optionally skip the integrity test when creating a new larry
· Add ability to compare (==, >, !=, etc) larrys with lists and tuples
· Documentation and unit tests
Breakage from la 0.2:
· lastrank and lastrank_decay methods combined into one method: lastrank
· Given shape (n,m) input, lastrank now returns shape (n,) instead of (n,1)
· geometric_mean now reduces input in the same way as lastrank (see above)
· #571813 Three larry methods crashed on 1d input
· #571737 skiprows missing from parameters section of the fromcsv doc string
· #571899 label indexing fails when larry is 3d and index is a tuple of len 2
· #571830 prod, cumprod, and cumsum did not return NaN for all-NaN input
· #572638 lastrank chokes on input with a shape tuple that contains zero
· #573240 Reduce methods give wrong output with shapes that contain zero
· #582579 la.afunc.nans: wrong output for str and object dtype
· #583596 assert_larry_equal crashed when comparing float larry to str larry
· #585694 cumsum and cumprod crashed on dtype=int