def run_simple_query(self, query):
(requested_columns, from_table, where_clause,
rest_of_query) = get_query_parts(query)
if where_clause != '':
where_exp = self.parse_where_clause(where_clause, from_table)
if not where_exp is None:
matches = where_exp.evaluate(self.session, "*")
if len(matches) == 0:
return OrderedDict([])
else:
try:
dink_query = "SELECT %s FROM %s" % (
','.join(requested_columns), from_table)
if matches != "*":
if from_table.startswith('samples'):
in_clause = "','".join(matches)
dink_query += " WHERE %s IN ('%s')" % (
self.get_partition_key(from_table), in_clause)
else:
in_clause = ",".join(map(str, matches))
dink_query += " WHERE %s IN (%s)" % (
self.get_partition_key(from_table), in_clause)
dink_query += " " + rest_of_query
self.session.row_factory = ordered_dict_factory
return self.session.execute(dink_query)
except cassandra.protocol.SyntaxException as e:
print "Cassandra error: {0}".format(e)
sys.exit("The query issued (%s) has a syntax error." % query)
def run_simple_query(self, query):
(requested_columns, from_table, where_clause, rest_of_query) = get_query_parts(query)
if where_clause != '':
where_exp = self.parse_where_clause(where_clause, from_table)
if not where_exp is None:
matches = where_exp.evaluate(self.session, "*")
if len(matches) == 0:
return OrderedDict([])
else:
try:
dink_query = "SELECT %s FROM %s" % (','.join(requested_columns), from_table)
if matches != "*":
if from_table.startswith('samples'):
in_clause = "','".join(matches)
dink_query += " WHERE %s IN ('%s')" % (self.get_partition_key(from_table), in_clause)
else:
in_clause = ",".join(map(str, matches))
dink_query += " WHERE %s IN (%s)" % (self.get_partition_key(from_table), in_clause)
dink_query += " " + rest_of_query
self.session.row_factory = ordered_dict_factory
return self.session.execute(dink_query)
except cassandra.protocol.SyntaxException as e:
print "Cassandra error: {0}".format(e)
sys.exit("The query issued (%s) has a syntax error." % query)
def run(self,
query,
gt_filter=None,
show_variant_samples=False,
variant_samples_delim=',',
predicates=None,
needs_genotypes=False,
needs_genes=False,
show_families=False,
test_mode=False,
needs_sample_names=False,
nr_cores=1,
start_time=-42,
use_header=False,
exp_id="Oink",
timeout=10.0,
batch_size=100):
"""
Execute a query against a Gemini database. The user may
specify:
1. (reqd.) an SQL `query`.
2. (opt.) a genotype filter.
"""
self.query = self.formatter.format_query(query).replace('==', '=')
self.gt_filter = gt_filter
#print self.query + '; gt-filter = %s \n' % gt_filter
self.nr_cores = nr_cores
self.start_time = start_time
self.use_header = use_header
self.exp_id = exp_id
self.timeout = timeout
self.batch_size = batch_size
if self._is_gt_filter_safe() is False:
sys.exit("ERROR: unsafe --gt-filter command.")
self.show_variant_samples = show_variant_samples
self.variant_samples_delim = variant_samples_delim
self.test_mode = test_mode
self.needs_genotypes = needs_genotypes
self.needs_vcf_columns = False
if self.formatter.name == 'vcf':
self.needs_vcf_columns = True
self.needs_sample_names = needs_sample_names
self.needs_genes = needs_genes
self.show_families = show_families
if predicates:
self.predicates += predicates
# make sure the SELECT columns are separated by a
# comma and a space. then tokenize by spaces.
self.query = self.query.replace(',', ', ')
self.query_pieces = self.query.split()
if not any(s.startswith("gt") for s in self.query_pieces) and \
not any(s.startswith("(gt") for s in self.query_pieces) and \
not any(".gt" in s for s in self.query_pieces):
if self.gt_filter is None:
self.query_type = "no-genotypes"
else:
self.gt_filter_exp = self._correct_genotype_filter()
self.query_type = "filter-genotypes"
else:
if self.gt_filter is None:
self.query_type = "select-genotypes"
else:
self.gt_filter_exp = self._correct_genotype_filter()
self.query_type = "filter-genotypes"
(self.requested_columns, self.from_table, where_clause,
self.rest_of_query) = get_query_parts(self.query)
self.extra_columns = []
if where_clause != '':
self.where_exp = self.parse_where_clause(where_clause,
self.from_table)
if not self.gt_filter is None:
self.where_exp = AND_expression(self.where_exp,
self.gt_filter_exp)
else:
if not self.gt_filter is None:
self.where_exp = self.gt_filter_exp
else:
self.where_exp = None
self._apply_query()
self.query_executed = True
def run(self, query, gt_filter=None, show_variant_samples=False,
variant_samples_delim=',', predicates=None,
needs_genotypes=False, needs_genes=False,
show_families=False, test_mode=False,
needs_sample_names=False, nr_cores = 1,
start_time = -42, use_header = False,
exp_id="Oink", timeout=10.0, batch_size = 100):
"""
Execute a query against a Gemini database. The user may
specify:
1. (reqd.) an SQL `query`.
2. (opt.) a genotype filter.
"""
self.query = self.formatter.format_query(query).replace('==','=')
self.gt_filter = gt_filter
#print self.query + '; gt-filter = %s \n' % gt_filter
self.nr_cores = nr_cores
self.start_time = start_time
self.use_header = use_header
self.exp_id = exp_id
self.timeout = timeout
self.batch_size = batch_size
if self._is_gt_filter_safe() is False:
sys.exit("ERROR: unsafe --gt-filter command.")
self.show_variant_samples = show_variant_samples
self.variant_samples_delim = variant_samples_delim
self.test_mode = test_mode
self.needs_genotypes = needs_genotypes
self.needs_vcf_columns = False
if self.formatter.name == 'vcf':
self.needs_vcf_columns = True
self.needs_sample_names = needs_sample_names
self.needs_genes = needs_genes
self.show_families = show_families
if predicates:
self.predicates += predicates
# make sure the SELECT columns are separated by a
# comma and a space. then tokenize by spaces.
self.query = self.query.replace(',', ', ')
self.query_pieces = self.query.split()
if not any(s.startswith("gt") for s in self.query_pieces) and \
not any(s.startswith("(gt") for s in self.query_pieces) and \
not any(".gt" in s for s in self.query_pieces):
if self.gt_filter is None:
self.query_type = "no-genotypes"
else:
self.gt_filter_exp = self._correct_genotype_filter()
self.query_type = "filter-genotypes"
else:
if self.gt_filter is None:
self.query_type = "select-genotypes"
else:
self.gt_filter_exp = self._correct_genotype_filter()
self.query_type = "filter-genotypes"
(self.requested_columns, self.from_table, where_clause, self.rest_of_query) = get_query_parts(self.query)
self.extra_columns = []
if where_clause != '':
self.where_exp = self.parse_where_clause(where_clause, self.from_table)
if not self.gt_filter is None:
self.where_exp = AND_expression(self.where_exp, self.gt_filter_exp)
else:
if not self.gt_filter is None:
self.where_exp = self.gt_filter_exp
else:
self.where_exp = None
self._apply_query()
self.query_executed = True