def importFromIterator(
outfile,
tablename,
iterator,
columns=None,
indices=None):
'''import data from an iterator into a database.
Arguments
---------
outfile : string
Output file name
tablename : string
Table name
iterator : iterator
Iterator to import data from. The iterator should
yield either list/tuples or dictionaries for each
row in the table.
columns : list
Column names. If not given, the assumption is that
iterator will dictionaries and column names are derived
from that.
indices : list
List of column names to add indices on.
'''
tmpfile = getTempFile(".")
if columns:
keys, values = list(zip(*list(columns.items())))
tmpfile.write("\t".join(values) + "\n")
for row in iterator:
if not columns:
keys = list(row[0].keys())
values = keys
columns = keys
tmpfile.write("\t".join(values) + "\n")
tmpfile.write("\t".join(str(row[x]) for x in keys) + "\n")
tmpfile.close()
if indices:
indices = " ".join("--add-index=%s" % x for x in indices)
else:
indices = ""
load(tmpfile.name,
outfile,
tablename=tablename,
options=indices)
os.unlink(tmpfile.name)
def importFromIterator(
outfile,
tablename,
iterator,
columns=None,
indices=None):
'''import data from an iterator into a database.
Arguments
---------
outfile : string
Output file name
tablename : string
Table name
iterator : iterator
Iterator to import data from. The iterator should
yield either list/tuples or dictionaries for each
row in the table.
columns : list
Column names. If not given, the assumption is that
iterator will dictionaries and column names are derived
from that.
indices : list
List of column names to add indices on.
'''
tmpfile = getTempFile(".")
if columns:
keys, values = list(zip(*list(columns.items())))
tmpfile.write("\t".join(values) + "\n")
for row in iterator:
if not columns:
keys = list(row[0].keys())
values = keys
columns = keys
tmpfile.write("\t".join(values) + "\n")
tmpfile.write("\t".join(str(row[x]) for x in keys) + "\n")
tmpfile.close()
if indices:
indices = " ".join("--add-index=%s" % x for x in indices)
else:
indices = ""
load(tmpfile.name,
outfile,
tablename=tablename,
options=indices)
os.unlink(tmpfile.name)
def buildJobScript(statement, job_memory, job_name):
'''build job script from statement.
returns (name_of_script, stdout_path, stderr_path)
'''
tmpfile = getTempFile(dir=PARAMS["workingdir"])
# disabled: -l -O expand_aliases\n" )
tmpfile.write("#!/bin/bash\n")
tmpfile.write(
'echo "%s : START -> %s" >> %s\n' %
(job_name, tmpfile.name, shellfile))
# disabled - problems with quoting
# tmpfile.write( '''echo 'statement=%s' >> %s\n''' %
# (shellquote(statement), shellfile) )
tmpfile.write("set | sed 's/^/%s : /' &>> %s\n" %
(job_name, shellfile))
# module list outputs to stderr, so merge stderr and stdout
tmpfile.write("module list 2>&1 | sed 's/^/%s: /' &>> %s\n" %
(job_name, shellfile))
tmpfile.write("hostname | sed 's/^/%s: /' &>> %s\n" %
(job_name, shellfile))
tmpfile.write("cat /proc/meminfo | sed 's/^/%s: /' &>> %s\n" %
(job_name, shellfile))
tmpfile.write(
'echo "%s : END -> %s" >> %s\n' %
(job_name, tmpfile.name, shellfile))
# restrict virtual memory
# Note that there are resources in SGE which could do this directly
# such as v_hmem.
# Note that limiting resident set sizes (RSS) with ulimit is not
# possible in newer kernels.
tmpfile.write("ulimit -v %i\n" % IOTools.human2bytes(job_memory))
tmpfile.write(
expandStatement(
statement,
ignore_pipe_errors=ignore_pipe_errors) + "\n")
tmpfile.close()
job_path = os.path.abspath(tmpfile.name)
stdout_path = job_path + ".stdout"
stderr_path = job_path + ".stderr"
os.chmod(job_path, stat.S_IRWXG | stat.S_IRWXU)
return (job_path, stdout_path, stderr_path)
