Search code examples
runzipglob

Sys.glob () within unzip ()

TLDNR: How do I use Sys.glob () within unzip ()?

I have multiple .zip files and I want to extract only one file from each archive.

For example, one of the archives contains the following files:

[1] "cmc-20150531.xml"     "cmc-20150531.xsd"     "cmc-20150531_cal.xml" "cmc-20150531_def.xml" "cmc-20150531_lab.xml"
[6] "cmc-20150531_pre.xml"

I want to extract the first file because it matches a pattern. In order to do that I use the following command:

unzip("zip-archive.zip", files=Sys.glob("[a-z][a-z][a-z][-][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][.][x][m][l]"))

However, the command doesn't work, and I don't know why. R just extracts all files in the archive.

On the other hand, the following command works:

unzip("zip-archive.zip", files="cmc-20150531.xml")

How do I use Sys.glob() within unzip()?

Solution

  • Sys.glob expands files that already exist. So the parameter to your unzip call will depend on what files are in your working directory.

    Perhaps you want to do unzip with list=TRUE to return the list of files in the zip first, and then use some pattern matching to select the files you want.

    See ?grep for info on matching strings with patterns. These patterns are "regular expressions" rather than "glob" expansions, but you should be able to work with that.

    Here's a concrete example:

    # whats in the zip?
files = unzip("c.zip", list=TRUE)$Name
files
[1] "l_spatial.dbf"    "l_spatial.shp"    "l_spatial.shx"    "ls_polys_bin.dbf"
[5] "ls_polys_bin.shp" "ls_polys_bin.shx" "rast_jan90.tif"  

# what files have "dbf" in them:
files[grepl("dbf",files)]
[1] "l_spatial.dbf"    "ls_polys_bin.dbf"

# extract just those:
unzip("c.zip", files=files[grepl("dbf",files)])

    The regular expression for your glob 

     "[a-z][a-z][a-z][-][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][.][x][m][l]"

    would be

     "^[a-z]{3}-[0-9]{8}\\.xml$"

    that's a match of start of string ("^"), 3 a-z (lower case only), a dash, eight digits, a dot (backslashes are needed, one because dot means "any one char" in regexps and another because R needs a backslash to escape a backslash), "xml", and the end of the string ("$").