Activity for Pascal Dupuis
-
Pascal Dupuis committed [86300f]
Cleaning the interface
-
Ensure the NA detection succeeds
-
an inputparser object is read-only
-
Merged local and remote changes
-
WIP -- try fixing deferencing issues
-
Fixed a few oddities
-
The row access with auto-fill is now OK
-
To simplify things, ridx is the same for all pages
-
Emptying single page seems OK
-
Generates an index for meta-data and another for array access
-
Added a field to flag lines as empty
-
A few modifications to handle emptied lines the right way
-
Pass ts to compute_tls to get graphs where time appears
-
Added back the 'Skip' parameter to the parser
-
svd: use 'econ' instead of 0 as second argument
-
A few stylistic changes to homogenize the whole program
-
Made arguments parsing MUCH stricter
-
Moved the regexp processing in df_regexp2name.m
-
Ensure asked_rownames is a vector
-
Perform searches using ismember
-
More robust detection of rownames and indexes
-
Added two files to test reading rownames and idx from csv file
-
Finished switching everything to the new df_name2idx interface
-
WIP: should be able to deal with x([1 1 1], :)
-
Adapted to new df_name2idx interface
-
In verbose mode, do not put df_matassign in a try-catch block
-
Repaired a broken merge
-
df_name2idx: index indexes instead of names internally
-
has_index: new boolean to use the first two columns from rownames and indexes
-
Adapted numel to new df_name2idx interface
-
Align the behavior on core func in data.cc
-
Ensure the added new lines indexes do not clash with existing ones
-
WIP: df_name2idx should be able to modify both elements of subs if input index is a df
-
Added tag dataframe-1.2.0 for changeset b5547db73d9e
-
The ':' magic should access ridx for row only
-
Pass the dataframe object to df_name2idx
-
On whole access, don't mess with idx
-
When rownames are present, sort them
-
Increased the tolerance for some tests
-
Updated the NEWS file
-
dataframe_test.m: detect non-failures where failures should occur
-
Made it back compatible with Octave 3.8
-
Removed extraneous indexes; all test should be OK
-
Updated license
-
Trying to debug a few oddities
-
dataframe-1.2
-
The md5 target is not implemented ???
-
Do-nothing script to ensure 'make release' works
-
Preparing for 1.2.0 release
-
Octave 4.2 support
-
VisualStudio2010 link error
-
Removed a leftover from the transition to field...
-
Solved a few oddities with the change of field ...
-
Querying a column name should return one value ...
-
Merged local and remote changes
-
Also handles the case where the input is a nons...
-
In case of multi-page data, collate everything ...
-
Merged Carnë changes
-
A dataframe can now be constructed from a struct
-
Better integration into Octave
-
Removed a typo
-
Placed two branches in sync
-
Updated all copyright statements
-
Initialize _ridx too
-
Calls the right wrapper
-
New functions
-
Implemented column iterate
-
Made the error message clearer
-
Many new tests, and an issue solved
-
Ensure _data and _rep have the right size
-
New functions testing if the content is boolean
-
More tests and sanity checks
-
Added a test for a selection over two columns
-
Ensure char type in input is reflected in all o...
-
Apply automatic broadcasting
-
df_allmeta: do not dump type 'char'
-
Preparing for 4.0 release
-
Various fixes in prevision of Octave 4.0.0
-
numel: also consider logical selectors
-
Implemented numel without calling subsref
-
when RHS is a matrix array, transform to a cellstr
-
subsasgn; better handling of cells third dim wh...
-
iterate on pages first
-
ischar logic must be sorted out; char: new func...
-
varargout can not be assigned in vector form at...
-
Overload some funcs to enable strmatch on dataf...
-
Enable the multi-argument call
-
Complex support added
1 >