NEWS
ncdfFlow 2.11.39
Enhancement
- update 'samples' slot of 'ncdfFlowList' class from 'character' to 'named integer' to speed up look up
- replace 'coerce' method with 'ncdfFlowList' constructor function
- update validity check of 'ncdfFlowList' class
- add 'subset.ncdfFlowList' and 'subset.ncdfFlowSet' S3 functions to subset the ncdfFlowSet/ncdfFlowList based on 'pData'
- wrap "[[" logic into c++ to speed it up
- add 'save_ncfs' and 'load_ncfs' functions to save/load a ncdfFlowSet object to/from disk.
ncdfFlow 1.5.32
- 1.add "file" argument to allow user to specify file path
2.add "rbind" method to allow combining more than two ncdfFlowSets once,
ncdfFlow 1.3.2
- 1.optimize [[ accessor by merging two routines into [[ method and removing some unnecessary check or subsetting
ncdfFlow 1.1.2
- 1.Using temporary directory instead of working directory to store cdf file in creating ncdfFlowSet from flowSet
2.allow for user specified path in ncdfFlowSet_sync method to save the cdf in different location other than original one
3.clone.ncdfFlowSet function:
-change argument name to avoid confusion:sNewNcFile-->isNew ;newNcFile-->fileName
-avoid copying the entire cdf repository when clone subsetted ncdfFlowSet
-fix the bug of inconsistent dimensions (sample*colnames) when create the new cdf file
4.check whether source file exist in read.ncdfFlowSet
5..writeSlice:
-allow for either flowFrame or matrix to be added by
-add sample name to the error message to help troubleshoot the problematic FCS file especially for loading large datasets
6.add isNew=FALSE to split method to allow for splitting into multipe cdf files for the sake of parallel computing
7.set compress=FALSE to disable compression mode of CDF
ncdfFlow 1.1.1
FEATURES
- -netCDF support for large data sets.
-centralized storage of flow data in 3-D matrix (sample*channel*event)
-fast data accessing,subsetting and splitting
-support all the related methods for flowSet
KNOWN ISSUES
- write meta data - ncdfFlow allows user to save the entire ncdfFlowSet object in ncdf file.
Currently the meta data is first serialized in R and stored as raw vector in cdf. It can fail
when the meta data size exceeds the limit of serialization function.