Skip to content

Instantly share code, notes, and snippets.

@NathanW2
Last active October 18, 2016 17:00
Show Gist options
  • Save NathanW2/4ab4c7eb81eccbf3825a7126a824c33e to your computer and use it in GitHub Desktop.
Save NathanW2/4ab4c7eb81eccbf3825a7126a824c33e to your computer and use it in GitHub Desktop.
query = From(iface.activeLayer()) \
.where('"Pit Type" = \'MH\'') \
.limit(10) \
.select("Pit Type", "SURVEY", "CATCHMENT", geometry=True)
query, layer = SQL("SELECT SURVEY, geometry FROM Pits GROUP BY Pit LIMIT 20", name="MyQuery")
query, layer = SQL("SELECT Pit, SURVEY, ST_UNION(ST_Buffer(geometry, 20)) as geometry FROM Pits GROUP BY Pit",
name="Buffers")
from qgis.core import QgsFeatureRequest
from PyQt4.QtCore import QUrl
class Query(object):
def __init__(self, layer):
self.rq = QgsFeatureRequest()
self.layer = layer
def where(self, expression):
self.rq.setFilterExpression(expression)
return self
def in_(self, *ids):
self.rq.setFilterFids(ids)
return self
def extent(self, rect):
self.rq.setFilterRect(rect)
return self
def no_geometry(self):
self.rq.setFlags(QgsFeatureRequest.NoGeometry)
return self
def limit(self, top):
self.rq.setLimit(top)
return self
def select(self, *fields, **kwargs):
if fields:
self.rq.setSubsetOfAttributes(fields, self.layer.pendingFields())
if "geometry" in kwargs and not kwargs['geometry']:
self.no_geometry()
return self.layer.getFeatures(self.rq)
def SQL(sql, name="querylayer"):
sql = "?query={0}&geometry=geometry".format(sql)
query = QUrl(sql).toEncoded()
layer = QgsVectorLayer(str(query), name, "virtual")
if not layer.dataProvider().error().isEmpty():
raise Exception(layer.dataProvider().error().message())
return From(layer).select(), layer
def From(layer):
return Query(layer)
@NathanW2
Copy link
Author

Used to query a layer in pyqgis in a friendly way without worrying about the underlying pyqgis API. Wraps QgsFetaureRequest for the method chaining style and virtual layers for the raw SQL style.

@tomchadwin
Copy link

Are you suggesting all three syntaxes be valid, or asking for opinions on which is best?

@ThomasG77
Copy link

Gain readability and expressibility with an SQL like syntax.
Performances costs? (because ORM are now less use to fully take advantage of the underlying DB)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment