b'Data trends Data trendsDEFN [continuation] ST=RECD,RT=[name]; {field definition}; {field definition}; etcDEFN 5 ST=RECORD,RT=DATA;FIDuCIAL:f12.1:NuLL=-999999.0,NAME=fiducialLines should split on the semi colons (;) to isolate the record type and each defined field. Then split on colons (:) and commas (,) to isolate values inside a field. There can be many fields on one line, but this is not recommended since one of the points of GDF2 is to be human readable. The above example would clearly become unreadable fast.Tim KeepingMisuse of the punctuation marks creates Associate Editor for geophysical havoc. Colons associate with the bare data management and analysisbones data type, and commas with the technical-standards@aseg.org.au optional parameters. The hierarchy of splitting relies on this.Splitting strings overNo DATA in the data fileIt is common for each line in the data ASEG GDF2 files file to start with DATA. Or not, and the difference is in the record type (RT). In the column I wrote for the FebruaryTherecord type setting directs which file issue of Preview (204), I identifiedthe line refers to.problems in ASEG GDF2 datasets thatDEFN ST=RECORD,RT=COMM;RT:A4confused my attempt at a parser. The standard is quite freeform andEach line in the description file (.des) will customisable, which results in astart with a four character record (A4), complex beast for all those tryingCOMM.to capture it. I have communicated with experts and re-read David PrattsDEFN ST=RECORD,RT=DATA;RT:A42003 publication, available from theEach line in the data file with start with Technical Standards page of the ASEGthe four character record DATA. But website, (https://www.aseg.org.au/ what about the mysterious null record technical/aseg-technical-standards).type?Light was shed and some ambiguities were set straight, while others becameDEFN ST=RECORD,RT=;RT:A4existential. As the bare minimum files in a GDF2 dataset are the .dfn and .dat, this defaults The definition file to RT=DATA.This first file created for a GDF2 datasetIs this all worth worrying about? A more is the definition file (.dfn) which makesmechanical outlook was given as since the data both computer and humanthis character field may or may not exist readable. It provides the field informationdepending on the supplier, and has no for data import that make GDF2 moreimpact on the data, some parser writers efficient to read than a single file suchhave learnt to detect but drop it. No as xyz or csv. It also cleverly providesharm is done. I hope this helps the other enough information to treat the text dataprogrammers trying to tame the GDF2.file (.dat) as a binary file.String splits and regular expressionsLets start with the generic syntax of each part of a .dfn line and a typical line.31 PREVIEW JuNE 2020'