Support globs and gzipped data with CSV FDW.

python/multicorn/csvfdw.py

@@ -6,30 +6,60 @@
 
 from . import ForeignDataWrapper
 from .utils import log_to_postgres
-from logging import WARNING
+from logging import INFO, WARNING, ERROR
 import csv
+import glob
+import gzip
+import itertools
+import os
+import re
 
 
 class CsvFdw(ForeignDataWrapper):
     """A foreign data wrapper for accessing csv files.
 
     Valid options:
-        - filename : full path to the csv file, which must be readable
-          by the user running postgresql (usually postgres)
-        - delimiter : the delimiter used between fields.
+        - filename : Full path to the csv file, which must be readable
+          by the user running postgresql (usually postgres).
+        - globs : Glob for recognizing files to process, which must be
+          readable by the user running postgresql (usually postgres). Multiple
+          globs may be passed, separated by ' // ' (space-slash-slash-space).
+          (UNIX paths can not meaningfully contain two slashes.)
+        - delimiter : The delimiter used between fields.
           Default: ","
+        - quotechar : The CSV quote character.
+          Default: '"'
+        - skip_header : The number of lines to skip.
+          Default: 0
+        - gzip : Look for GZip file magic and decompress files if found.
+          Default: true
     """
 
     def __init__(self, fdw_options, fdw_columns):
         super(CsvFdw, self).__init__(fdw_options, fdw_columns)
-        self.filename = fdw_options["filename"]
+        self.filename = fdw_options.get("filename")
         self.delimiter = fdw_options.get("delimiter", ",")
         self.quotechar = fdw_options.get("quotechar", '"')
-        self.skip_header = int(fdw_options.get('skip_header', 0))
+        self.skip_header = int(fdw_options.get("skip_header", 0))
         self.columns = fdw_columns
+        self.globs = set(fdw_options.get("globs", "").split(" // "))
+        self.globs.discard("")
+        if not (bool(self.globs) ^ bool(self.filename)):
+            log_to_postgres("Please set either 'filename' or 'globs'.", ERROR)
+        try:
+            self.gzip = {"true":  True,
+                         "false": False}[fdw_options.get("gzip", "true")]
+        except KeyError:
+            log_to_postgres("Please set 'gzip' as 'true' or 'false'.", ERROR)
+
 
     def execute(self, quals, columns):
-        with open(self.filename) as stream:
+        generators = (self.load_file(path, columns) for path in self.paths())
+        return itertools.chain(*generators)
+
+    def load_file(self, path, columns):
+        open_function = self.opener(path)
+        with open_function(path) as stream:
             reader = csv.reader(stream, delimiter=self.delimiter)
             count = 0
             checked = False
@@ -40,10 +70,24 @@ def execute(self, quals, columns):
                         # appropriate length
                         checked = True
                         if len(line) > len(self.columns):
-                            log_to_postgres("There are more columns than "
-                                            "defined in the table", WARNING)
+                            log_to_postgres("More columns than defined in "
+                                            "table: %s" % path, WARNING)
                         if len(line) < len(self.columns):
-                            log_to_postgres("There are less columns than "
-                                            "defined in the table", WARNING)
+                            log_to_postgres("Fewer columns than defined in "
+                                            "table: %s" % path, WARNING)
                     yield line[:len(self.columns)]
                 count += 1
+
+    def opener(self, path):
+        with open(path) as stream:
+            if self.gzip and stream.read(2) == b'\x1F\x8b':
+                log_to_postgres("Reading CSV with gzip: %s" % path, INFO)
+                return gzip.open
+            return open
+
+    def paths(self):
+        if self.globs:
+            joined = itertools.chain(*(glob.iglob(g) for g in self.globs))
+            return (p for p in joined if os.path.isfile(p))
+        else:
+            return [self.filename]