makeDB: Create a seismic Waveform Database

In RSEIS: Seismic Time Series Analysis Tools

Create a seismic Waveform Database

Description

Create a seismic Waveform Database

Usage

makeDB(path=".", pattern="R", dirs="", kind = 1,
Iendian=1, BIGLONG=FALSE)
FmakeDB(LF2, kind =1, Iendian=1, BIGLONG=FALSE)

Arguments

path	character, Path to directory where files and directories exist
pattern	character, pattern for listing of files
dirs	character, vector of directories to be scanned
kind	kind of data: RDS=-1, R(DAT)=0, segy=1; sac=2
Iendian	default=1, Endian-ness of the data: 1,2,3: "little", "big", "swap". Default = 1 (little)
BIGLONG	logical, TRUE means long=8 bytes
LF2	list of files

Details

The files are typically located in a directory structure created by programs like ref2segy, a PASSCAL program for downloading data in the field. Each file contains one seismogram, with a header. makeDB reads in all the headers and creates a list of meta-data for later use in RSEIS.

"kind" can be numeric or character: options are 'RDS', 'RDATA', 'SEGY', 'SAC', corresponding to (-1, 0, 1, 2).

Uses readBin to extract data in SAC format. user must know what kind of machine the data was created on for I/O purposes.

If data was created on a little endian machine but is being read on big endian machine, need to call the endian "swap" for swapping.

If data was created on a machine with LONG=4 bytes, be sure to call the program with BIGLONG=FALSE.

If the base directory, or the subdirectories, contain files that are not seismic data then care must be taken. Perhaps use FmakeDB to explicitly names the files for the DataBase.

If using FmakeDB a simple vector of files (full path names) should be provided.

The origin year, used for getting the Epoch year, is stored as attribute origyr.

Value

list:

fn	file name
yr	year
jd	julian day
hr	hour
mi	minute
sec	second
dur	duration, seconds
t1	time 1 in Epoch days
t2	time 2 in Epoch days
sta	station name
comp	component name
dt	sample rate, seconds

Note

Epoch times are used to accomodate problems where julian days cross year end boundaries, so that day 366 comes before day 1 of the next year.

The origyr, kind , Iendian, BIGLONG are stored as attributes in the Database.

Author(s)

Jonathan M. Lees<[e-mail]>

See Also

setupDB, Mine.seis , getseis24, plotseis24, EPOCHday, swig

Examples

##########  to illustrate, we make a set of individual seismograms
data(GH)
L1 = length(GH$JSTR)
DD = data.frame(GH$info)

GIVE = vector(mode='list')

for(i in 1:L1)
{
AA = DD[i,]
GIVE[[i]] = list(fn = AA$fn, sta =GH$STNS[i] , comp = GH$COMP[i], 
            dt = AA$dt, DATTIM = AA, N = AA$n1, units = NA, 
            coords = NA, amp = GH$JSTR[[i]]  )
}

###########  save the seismic data in a temporary directory
####   each trace in a separate file
tdir = tempdir()
for(i in 1:length(GIVE) )
{
    sig = GIVE[[i]]
  d1 =  dateStamp(sig$DATTIM, sep='_')
   nam1 =  paste(d1,sig$sta, sig$comp, sep='_')
    nam2 = paste0(nam1, '.RDS')
    nam3 = paste(tdir, nam2, sep='/')
    saveRDS(file=nam3, sig) 
    }

########################  Now  read files and make the DataBase:
LF  = list.files(path=tdir,pattern='.RDS',  full.names=TRUE)
DB = FmakeDB(LF, kind=-1)

RSEIS documentation built on Sept. 13, 2024, 1:09 a.m.