def to_ruby(self, schema):
if not self.terms:
return NULL.to_ruby(schema)
v = self.terms[-1]
acc = FirstOp("first", v).partial_eval().to_ruby(schema)
for v in reversed(self.terms[:-1]):
m = v.missing().partial_eval()
e = NotOp("not", m).partial_eval().to_ruby(schema)
r = FirstOp("first", v).partial_eval().to_ruby(schema)
if r.miss is TRUE:
continue
elif r.miss is FALSE:
acc = r
continue
elif acc.type == r.type:
new_type = r.type
elif acc.type == NUMBER and r.type == INTEGER:
new_type = NUMBER
elif acc.type == INTEGER and r.type == NUMBER:
new_type = NUMBER
else:
new_type = OBJECT
acc = Ruby(miss=AndOp("and", [acc.miss, m]).partial_eval(),
type=new_type,
expr="(" + e.expr + ") ? (" + r.expr + ") : (" + acc.expr +
")",
frum=self)
return acc
def to_ruby(self, schema):
if self.var == ".":
return "_source"
else:
if self.var == "_id":
return Ruby(
type=STRING,
expr=
'doc["_uid"].value.substring(doc["_uid"].value.indexOf(\'#\')+1)',
frum=self)
columns = schema.values(self.var)
acc = []
for c in columns:
varname = c.es_column
frum = Variable(c.es_column)
q = quote(varname)
acc.append(
Ruby(miss=frum.missing(),
type=c.type,
expr="doc[" + q + "].values",
frum=frum,
many=True))
if len(acc) == 0:
return NULL.to_ruby(schema)
elif len(acc) == 1:
return acc[0]
else:
return CoalesceOp("coalesce", acc).to_ruby(schema)
def to_es_script(self, schema):
if not self.terms:
return NULL.to_es_script(schema)
v = self.terms[-1]
acc = FirstOp("first", v).partial_eval().to_es_script(schema)
for v in reversed(self.terms[:-1]):
m = v.missing().partial_eval()
e = NotOp("not", m).partial_eval().to_es_script(schema)
r = FirstOp("first", v).partial_eval().to_es_script(schema)
if r.miss is TRUE:
continue
elif r.miss is FALSE:
acc = r
continue
elif acc.type == r.type:
new_type = r.type
elif acc.type == NUMBER and r.type == INTEGER:
new_type = NUMBER
elif acc.type == INTEGER and r.type == NUMBER:
new_type = NUMBER
else:
new_type = OBJECT
acc = EsScript(
miss=AndOp("and", [acc.miss, m]).partial_eval(),
type=new_type,
expr="(" + e.expr + ") ? (" + r.expr + ") : (" + acc.expr + ")",
frum=self
)
return acc
def to_es_script(self, schema, not_null=False, boolean=False, many=True):
if not self.terms:
return NULL.to_es_script(schema)
# acc.miss WILL SAY IF THIS COALESCE RETURNS NULL,
# acc.expr WILL ASSUMED TO BE A VALUE, SO THE LAST TERM IS ASSUMED NOT NULL
v = self.terms[-1]
acc = FirstOp(v).partial_eval().to_es_script(schema)
for v in reversed(self.terms[:-1]):
m = v.missing().partial_eval()
e = NotOp(m).partial_eval().to_es_script(schema)
r = FirstOp(v).partial_eval().to_es_script(schema)
if r.miss is TRUE:
continue
elif r.miss is FALSE:
acc = r
continue
elif acc.type == r.type or acc.type == IS_NULL:
new_type = r.type
elif acc.type == NUMBER and r.type == INTEGER:
new_type = NUMBER
elif acc.type == INTEGER and r.type == NUMBER:
new_type = NUMBER
else:
new_type = OBJECT
acc = EsScript(
miss=AndOp([acc.miss, m]).partial_eval(),
type=new_type,
expr="(" + e.expr + ") ? (" + r.expr + ") : (" + acc.expr +
")",
frum=self,
schema=schema,
)
return acc
def to_es_script(self, schema, not_null=False, boolean=False, many=True):
if not self.terms:
return NULL.to_es_script(schema)
# acc.miss WILL SAY IF THIS COALESCE RETURNS NULL,
# acc.expr WILL ASSUMED TO BE A VALUE, SO THE LAST TERM IS ASSUMED NOT NULL
v = self.terms[-1]
acc = FirstOp(v).partial_eval().to_es_script(schema)
for v in reversed(self.terms[:-1]):
m = v.missing().partial_eval()
e = NotOp(m).partial_eval().to_es_script(schema)
r = FirstOp(v).partial_eval().to_es_script(schema)
if r.miss is TRUE:
continue
elif r.miss is FALSE:
acc = r
continue
elif acc.type == r.type or acc.type == IS_NULL:
new_type = r.type
elif acc.type == NUMBER and r.type == INTEGER:
new_type = NUMBER
elif acc.type == INTEGER and r.type == NUMBER:
new_type = NUMBER
else:
new_type = OBJECT
acc = EsScript(
miss=AndOp([acc.miss, m]).partial_eval(),
type=new_type,
expr="(" + e.expr + ") ? (" + r.expr + ") : (" + acc.expr + ")",
frum=self,
schema=schema,
)
return acc
def _convert(v):
if v is None:
return NULL.to_es_script(schema)
if v is True:
return EsScript(
type=BOOLEAN,
expr="true",
frum=self
)
if v is False:
return EsScript(
type=BOOLEAN,
expr="false",
frum=self
)
if isinstance(v, text_type):
return EsScript(
type=STRING,
expr=quote(v),
frum=self
)
if isinstance(v, int):
return EsScript(
type=INTEGER,
expr=text_type(v),
frum=self
)
if isinstance(v, float):
return EsScript(
type=NUMBER,
expr=text_type(v),
frum=self
)
if isinstance(v, dict):
return EsScript(
type=OBJECT,
expr="[" + ", ".join(quote(k) + ": " + _convert(vv) for k, vv in v.items()) + "]",
frum=self
)
if isinstance(v, (list, tuple)):
return EsScript(
type=OBJECT,
expr="[" + ", ".join(_convert(vv).expr for vv in v) + "]",
frum=self
)
if isinstance(v, Date):
return EsScript(
type=NUMBER,
expr=text_type(v.unix),
frum=self
)
def to_es_script(self, schema, not_null=False, boolean=False, many=True):
if self.var == ".":
return EsScript(type=OBJECT, expr="_source", frum=self)
else:
if self.var == "_id":
return EsScript(
type=STRING,
expr='doc["_uid"].value.substring(doc["_uid"].value.indexOf(\'#\')+1)',
frum=self,
schema=schema,
)
columns = schema.values(self.var)
acc = []
for c in columns:
varname = c.es_column
frum = Variable(c.es_column)
q = quote(varname)
if many:
acc.append(
EsScript(
miss=frum.missing(),
type=c.jx_type,
expr="doc[" + q + "].values"
if c.jx_type != BOOLEAN
else "doc[" + q + "].value",
frum=frum,
schema=schema,
many=c.jx_type != BOOLEAN,
)
)
else:
acc.append(
EsScript(
miss=frum.missing(),
type=c.jx_type,
expr="doc[" + q + "].value"
if c.jx_type != BOOLEAN
else "doc[" + q + "].value",
frum=frum,
schema=schema,
many=True,
)
)
if len(acc) == 0:
return NULL.to_es_script(schema)
elif len(acc) == 1:
return acc[0]
else:
return CoalesceOp(acc).to_es_script(schema)
def to_es_script(self, schema, not_null=False, boolean=False, many=True):
if self.var == ".":
return EsScript(type=OBJECT, expr="_source", frum=self)
else:
if self.var == "_id":
return EsScript(
type=STRING,
expr='doc["_uid"].value.substring(doc["_uid"].value.indexOf(\'#\')+1)',
frum=self,
schema=schema,
)
columns = schema.values(self.var)
acc = []
for c in columns:
varname = c.es_column
frum = Variable(c.es_column)
q = quote(varname)
if many:
acc.append(
EsScript(
miss=frum.missing(),
type=c.jx_type,
expr="doc[" + q + "].values"
if c.jx_type != BOOLEAN
else "doc[" + q + "].value",
frum=frum,
schema=schema,
many=c.jx_type != BOOLEAN,
)
)
else:
acc.append(
EsScript(
miss=frum.missing(),
type=c.jx_type,
expr="doc[" + q + "].value"
if c.jx_type != BOOLEAN
else "doc[" + q + "].value",
frum=frum,
schema=schema,
many=True,
)
)
if len(acc) == 0:
return NULL.to_es_script(schema)
elif len(acc) == 1:
return acc[0]
else:
return CoalesceOp(acc).to_es_script(schema)
def to_bq(self, schema, not_null=False, boolean=False):
sql = []
if not self.terms:
return NULL.to_bq(schema)
for term in self.terms:
sql.append(BQLang[term].to_bq(schema))
return BQLScript(data_type=merge_json_type(*(t.type for t in sql)),
expr=sql_call("COALESCE", *sql),
frum=self,
miss=AndOp([MissingOp(t)
for t in self.terms]).partial_eval(),
many=False,
schema=schema)