search

SEL-Columbia / bamboo

Dynamic data analysis over the web. The logic to your data dashboards.
http://bamboo.io
BSD 3-Clause "New" or "Revised" License
156 stars 20 forks source link

erasing data sets! #350

Closed mberg closed 11 years ago

mberg commented 11 years ago

This erases the data set!

curl -X POST -F csv_file=@/home/mberg/Documents/bamboo-examples/data/water_points.csv http://bamboo.io/datasets

curl -X POST -d "name=protected_waterpoint&formula=water_source_type in [\"borehole\"]" http://bamboo.io/calculations/d997e6877f8a4d65b5eae6a3995c6773
mberg commented 11 years ago

This is what calculations show.

[ { formula: "water_source_type in ["borehole"]", state: "pending", group: null, name: "protected_waterpoint" } ]

mberg commented 11 years ago

This too:

curl -X POST -d "name=big_poppa&formula=community_pop>5000" http://bamboo.io/calculations/ef6995b3b2894e969248122ac586613a

pld commented 11 years ago

This is related to the format of the dataset that is uploaded. The water_points.csv has duplicate column names (at least the one in the github repo) and pandas does not like this.

For the time being #354 will silently fail but not delete the dataset, an improvement. I'm going to add an issue for uniquifying columns on upload.

pld commented 11 years ago

Closing, since the deletion issue is fixed. But still any operations on datasets with duplicate column names will result in unexpected behavior, issue for this is #355.