Write Compressed Data to MDF Files
This example shows how to write data to MDF files using compression, enhancing storage efficiency.
The MAT file MDFTimetables.mat used in this example contains two timetables.
Timetable
TT1has variables of numeric data types, including uint8, uint16, uint32, uint64, int8, int16, int32, int64, single, and double.Timetable
TT2has variables of string and byte array data types, including both fixed-length and variable-length data.
Introduction
According to the ASAM MDF standard, the Data Zipped Block (DZBLOCK), introduced in MDF version 4.1, allows for the compression of various data blocks within MDF files, which is particularly useful when disk space is limited. By compressing signal data during the writing process, it effectively reduces file size. However, for very small data sections, the compression overhead might not provide a size advantage. Tools for reading can decompress the data to access the original data.
ASAM MDF V4.2.0 supports the following DZBLOCK compression algorithms:
dz_zip_type = 0: Deflate
dz_zip_type = 1: Transposition + Deflate
The mdfWrite function by default does not write with compression. To write with compression, the CompressionType name-value argument must be set to a value other than "none".
The CompressionType name-value argument for mdfWrite controls the compression algorithm used when writing data to the MDF file:
"
none" (default): Writes uncompressed data."
deflate": Writes compressed data using the Deflate algorithm."
transposeDeflate": Writes compressed data using the Transpose-Deflate algorithm.
The CompressionLevel name-value argument accepts a numerical value and controls the level of compression, depending on the CompressionType:
For "
none": the default and only acceptable value is 0.For "
deflate" and "transposeDeflate": the default value is 6 and acceptable values are from 0 (no compression) to 9 (maximum compression). Increasing the compression level will result in longer writing times.
Load Timetables into Workspace
Load the timetable variables, TT1 and TT2, from MDFTimetables.mat into the workspace.
load("MDFTimetables.mat")Write Compressed Data Directly to a New MDF File
Use the mdfWrite function with the target MDF file name, the timetable variable TT1, and the "deflate" compression type specified as input arguments. Since the CompressionLevel name-value argument is not specified, the default compression level of the specified compression type will be used, in this case the default level for "deflate" is 6. Data in TT1 is written to the first available channel group index of the file, which is channel group 1 since CompressedSignalData.mf4 is a newly created file.
mdfWrite("CompressedSignalData.mf4", TT1, CompressionType="deflate")
Warning: Specified target file did not exist. Created file: C:\ExampleManager9a362a\user.Bdoc.Compression_Example\vnt-ex52002539\CompressedSignalData.mf4
Writing data with compression is channel group based, meaning that a different compression type can be used for writing each channel group. Call mdfWrite again to write TT2 to the next available channel group index 2, this time writing with a compression type of "transposeDeflate" and a compression level of 9 (maximum compression, slowest speed).
mdfWrite("CompressedSignalData.mf4", TT2, CompressionType="transposeDeflate", CompressionLevel=9)
Examine the Data
To verify that data was successfully written to file, read data from both channel groups using mdfRead and examine the results.
data = mdfRead("CompressedSignalData.mf4")data=2×1 cell array
100×10 timetable
10×4 timetable
chanGrp1Data = data{1}chanGrp1Data=100×10 timetable
0 sec 0 200 600 1400 -99 -198 -396 -794 -9.8000 -19.6000
0.1 sec 2 204 608 1416 -97 -194 -388 -778 -9.6000 -19.2000
0.2 sec 4 208 616 1432 -95 -190 -380 -762 -9.4000 -18.8000
0.3 sec 6 212 624 1448 -93 -186 -372 -746 -9.2000 -18.4000
0.4 sec 8 216 632 1464 -91 -182 -364 -730 -9 -18
0.5 sec 10 220 640 1480 -89 -178 -356 -714 -8.8000 -17.6000
0.6 sec 12 224 648 1496 -87 -174 -348 -698 -8.6000 -17.2000
0.7 sec 14 228 656 1512 -85 -170 -340 -682 -8.4000 -16.8000
0.8 sec 16 232 664 1528 -83 -166 -332 -666 -8.2000 -16.4000
0.9 sec 18 236 672 1544 -81 -162 -324 -650 -8 -16
1 sec 20 240 680 1560 -79 -158 -316 -634 -7.8000 -15.6000
1.1 sec 22 244 688 1576 -77 -154 -308 -618 -7.6000 -15.2000
1.2 sec 24 248 696 1592 -75 -150 -300 -602 -7.4000 -14.8000
1.3 sec 26 252 704 1608 -73 -146 -292 -586 -7.2000 -14.4000
⋮
chanGrp2Data = data{2}chanGrp2Data=10×4 timetable
0 sec "abcd" [255,255,255,255,255] "zero" [255,255]
1 sec "efgh" [18,35,52,69,86] "one" [18,35,52,69,86]
2 sec "ijkl" [0,1,2,3,4] "two" [0,1,2,3,4,5,6,7]
3 sec "mnop" [4,3,2,1,0] "three" [4,3,2,1,0]
4 sec "qrst" [255,254,253,252,251] "four" [253,252,251]
5 sec "uvwx" [250,249,248,247,246] "five" [250,249,248,247,246]
6 sec "yzAB" [245,244,243,242,241] "six" [245,244,243,242,241]
7 sec "CDEF" [240,239,238,237,236] "seven" [240,238,236]
8 sec "GHIJ" [235,234,233,232,231] "eight" [235,233,231]
9 sec "KLMN" [255,255,255,255,255] "nine" [255,255,255,255,255]
Compare File Size to MDF File with Uncompressed Signal Data
Use the mdfWrite function to write TT1 and TT2 without compression to a new MDF file called UncompressedSignalData.mf4.
mdfWrite("UncompressedSignalData.mf4", TT1)Warning: Specified target file did not exist. Created file: C:\ExampleManager9a362a\user.Bdoc.Compression_Example\vnt-ex52002539\UncompressedSignalData.mf4
mdfWrite("UncompressedSignalData.mf4", TT2)Use the dir function to get the file size in bytes of both MDF files and compare.
compressedFileSizeBytes = dir("CompressedSignalData.mf4").bytescompressedFileSizeBytes = 11864
uncompressedFileSizeBytes = dir("UncompressedSignalData.mf4").bytesuncompressedFileSizeBytes = 14640
reductionPercentage = ((uncompressedFileSizeBytes - compressedFileSizeBytes) / uncompressedFileSizeBytes) * 100;
fprintf("%.2f%% reduction in MDF file size when writing signal data with compression.\n", reductionPercentage)18.96% reduction in MDF file size when writing signal data with compression.
