Commit Graph

479 Commits

Author SHA1 Message Date
8ce4f264aa When enlarging a column, the function doesn't try anymore to keep the
mapped region at the same pointer (never works), and unmap/remap
instead.
2015-11-10 13:18:36 +01:00
d885eb48ff The header size when creating a column is now calculated according to
the size of the header structure and the page size of the platform.
2015-11-10 13:09:30 +01:00
c4b7e579cf Comments in column headers are now working. 2015-11-10 10:56:45 +01:00
f6ec8ba963 The header size is now directly read in the file when a column or an
array is opened.
2015-11-09 17:50:32 +01:00
01bfc14503 The data size in bytes is now stored in the header of a column. 2015-11-09 15:55:00 +01:00
65c1b1e8b2 Minor changes to make the creation of files and directories cleaner 2015-11-09 15:22:01 +01:00
b37bd8f21c File descriptors for dms, column and array directories are now stored in
structures.
2015-11-09 15:06:02 +01:00
05e3956a0c Minor changes in code to improve readability (freeing some character
strings earlier)
2015-11-09 11:22:51 +01:00
9b066f4327 Major update: obiarrays with columns containing indices referring to
character strings.
2015-11-06 17:55:15 +01:00
456551ffeb obi arrays that don't work because of cython bug passing wrong pointers 2015-11-03 14:22:00 +01:00
ecb9d97adb Reorganized the code to have less functions, and the functions to get
and format the creation date of a column are now working.
2015-10-15 15:12:45 +02:00
0eaa5aa784 Major changes : new cython subclasses to handle columns with multiple
elements per line in a more efficient way + now elements_names are
passed as a list + new function to recover only the header of a column
2015-10-14 18:05:34 +02:00
6877fc4892 Fixed a critical bug where values were initialized to NA at the wrong
location when there was multiple elements per line
2015-10-12 17:54:36 +02:00
fc8bf16769 Fixed a critical bug in the computation of the new number of lines of a
column when truncating
2015-10-09 13:49:48 +02:00
e114a3c9cb fixed a critical bug where data size was not calculated correctly and
column directory is now closed when column is closed
2015-10-09 10:25:40 +02:00
ee4c513fd4 Fixed a bug where cloning a column would fail if the data was empty 2015-10-08 10:36:02 +02:00
c013e6ad33 fixed typo in doxygen doc 2015-10-08 10:33:19 +02:00
6ced3c4896 new functions to get the creation date of a column 2015-10-02 13:47:53 +02:00
4b8bf41a71 closes #13, obi_errno is initialized to 0 2015-10-02 13:46:34 +02:00
c59a244e9d Fixed little typo 2015-09-30 12:07:13 +02:00
4b7f2d268b Doxygen documentation corrected and completed. 2015-09-30 12:03:46 +02:00
45af8396b8 minor change in variable name for better coherence 2015-09-28 13:52:41 +02:00
b0570ee486 fixed permissions when creating files 2015-09-22 15:52:07 +02:00
72105b2aed fixed the computation of the data size when cloning a column 2015-09-22 10:27:17 +02:00
9ab8a42340 fixed the mkdir permission mode that forced the user to use sudo to
write in the created directories
2015-09-21 15:48:02 +02:00
7b606c0477 Column files now always have a size that is a multiple of the page size,
and the function that enlarges mapped column files tries mapping on next
byte before unmapping/mapping again
2015-09-21 15:42:29 +02:00
7d7dbb1bf9 Handling of single char columns 2015-09-18 15:49:15 +02:00
2698022aaf New OBIDMS method to list the columns of an OBIDMS 2015-09-15 17:09:31 +02:00
5c3bc03bd2 Added the handling of errors with the functions to get a value in a
column
2015-09-14 17:04:29 +02:00
f2231770f1 If an error occurs when cloning a column, the new cloned file is now
deleted
2015-09-02 13:06:21 +02:00
52e3f2ce4a Columns that are cloned are now closed after cloning 2015-09-02 10:36:00 +02:00
f31d8983bb Columns are now enlarged if needed when setting a value 2015-09-01 17:38:08 +02:00
3b7536c0df The number of lines used is updated in the header when a value is set,
and the iteration on the values of a column is done on the number of
lines used (instead of on the total line count).
2015-08-26 17:25:15 +02:00
34ade161de added an obidebug() line 2015-08-26 17:05:37 +02:00
48d10ea17c Functions to truncate and/or close a column 2015-08-26 17:01:54 +02:00
95dbeb25a0 Added functions to get and set values in columns using the element name
(for columns with lines made of vectors of elements), for all data types
2015-08-26 15:05:23 +02:00
9d91e907e5 When a column is created, its data is initialized to the NA value of the
obitype
2015-08-26 12:00:38 +02:00
3f81dc0173 Changed the variable name 'nb_elements' to 'nb_lines' for better clarity 2015-08-26 10:52:19 +02:00
8ff3488926 Fix: The number of lines used is set when cloning a column with its data 2015-08-26 10:38:07 +02:00
611699252e Files and functions for columns with the data type OBIChar_t, working
but using char* (needs to be changed)
2015-08-26 10:31:56 +02:00
9ad31fddff Added the possibility to clone a column, with or without its data 2015-08-26 10:29:07 +02:00
eaca9c85f1 Files and functions for OBIDMS columns with the data type OBI_BOOL 2015-08-11 11:00:59 +02:00
bc2ca89088 Basis for OBIDMS columns with the type OBI_IDX 2015-08-10 16:30:55 +02:00
3b3267db9a Removed a useless #include 2015-08-10 16:05:34 +02:00
d53c16cf3e Functions and files for OBIDMS columns with the type OBI_FLOAT 2015-08-10 16:04:53 +02:00
1fe2e36d5d Fixed a critical bug where the elements names were not stored correctly 2015-08-03 16:11:55 +02:00
ab44e32afb More efficient obi_column_set_int() and obi_column_get_int() 2015-08-03 15:26:28 +02:00
41f627091f Error handling: obidebug() with message for all errors, and removed
commands closing directories when an error occurred (creating more
errors).
2015-08-03 15:10:39 +02:00
1e01c9059c functions to get data type as a character string 2015-08-03 11:46:21 +02:00
5f62cd8526 C sources modified to add the handling of OBIDM columns with the type
OBI_INT, and the handling of multiple types in general
2015-07-31 18:03:48 +02:00