issues

Hi, datasette version: "0.26.2" extensions: spatialite: "4.4.0-RC0" sqlite version: "3.22.0"

I have a table of airports with latitude and longitude columns. I've added spatialite (with KNN support). After creating the db using csvs-to-sqlit, I run these commands to setup the spatialite tables:

``` conn.execute('SELECT InitSpatialMetadata(1)')

conn.execute("SELECT AddGeometryColumn('airports', 'point_geom', 4326, 'POINT', 2);")

conn.execute('''UPDATE airports SET point_geom = GeomFromText('POINT('||"longitude"||' '||"latitude"||')',4326);''')

conn.execute("SELECT CreateSpatialIndex('airports', 'point_geom');") ```

I'm attempting to create a canned query and have this in my metadata.json file: "find_airports_nearest_to_point":{ "sql":"SELECT a.pos AS rank, b.id, b.name, b.country, b.latitude AS latitude, b.longitude AS longitude, a.distance / 1000.0 AS dist_km FROM KNN AS a JOIN airports AS b ON (b.rowid = a.fid) WHERE f_table_name = \"airports\" AND ref_geometry = MakePoint( :Long , :Lat ) AND max_items = 10;"} which doesn't seem to perform the templating of the name parameters correctly and I get no results.

Have also tired: MakePoint( || :Long || , || :Lat || ) which returns this error: near "||": syntax error

However I cannot seem to find the correct combination of named parameter syntax (:Lat) or sqlite concatenation operator to make it work. Any ideas if using named parameters inside functions is supported?

Thanks Darren

Hi,

I've been experimenting with SQLite reading from huge datasets using this excellent Parquet extension from @cldellow. https://cldellow.com/2018/06/22/sqlite-parquet-vtable.html https://github.com/cldellow/sqlite-parquet-vtable

This works really well, but I was keen to see if I could combine datasette with this. Having previously experimented with the spatialite extension I knew that datasette supports loading extensions in the underlying sqlite instance. However I hit a blocker as the current design only allows SELECT statements to be executed and so I am unable to execute the crucial

CREATE VIRTUAL TABLE .........

command that is required to load the data from the parquet file into the table.

It seems like this would be a simple-ish change, but I don't know enough about the architecture of datasette to start implementing this myself? Could this be done as a datasette plugin? or would this require more fundamental changes at initialisation time?

My thoughts are that something at init time could detect that the user was loading a .parquet file and then switch to a mode were it loads that via the "CREATE VIRTUAL TABLE..." rather than loading the .db file in the default case??

I'm happy to contribute code and testing, I just need some pointers on the best approach.

Thanks Darren

``` $ github-to-sqlite repos github.db $ github-to-sqlite starred github.db dazzag24

Traceback (most recent call last): File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/bin/github-to-sqlite", line 10, in <module> sys.exit(cli()) File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/click/core.py", line 764, in call return self.main(args, kwargs) File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/click/core.py", line 717, in main rv = self.invoke(ctx) File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/click/core.py", line 1137, in invoke return _process_result(sub_ctx.command.invoke(sub_ctx)) File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/click/core.py", line 956, in invoke return ctx.invoke(self.callback, ctx.params) File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/click/core.py", line 555, in invoke return callback(args, **kwargs) File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/github_to_sqlite/cli.py", line 106, in starred utils.save_stars(db, user, stars) File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/github_to_sqlite/utils.py", line 177, in save_stars user_id = save_user(db, user) File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/github_to_sqlite/utils.py", line 61, in save_user return db["users"].upsert(to_save, pk="id").last_pk File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/sqlite_utils/db.py", line 1067, in upsert extracts=extracts, File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/sqlite_utils/db.py", line 916, in insert extracts=extracts, File "/home/darreng/.virtualenvs/dogsheep-d2PjdrD7/lib/python3.6/site-packages/sqlite_utils/db.py", line 1024, in insert_all result = self.db.conn.execute(sql, values) sqlite3.OperationalError: table users has no column named bio

```

``` $ pipenv graph github-to-sqlite==0.4 - requests [required: Any, installed: 2.22.0] - certifi [required: >=2017.4.17, installed: 2019.9.11] - chardet [required: >=3.0.2,<3.1.0, installed: 3.0.4] - idna [required: >=2.5,<2.9, installed: 2.8] - urllib3 [required: >=1.21.1,<1.26,!=1.25.1,!=1.25.0, installed: 1.25.6] - sqlite-utils [required: ~=1.11, installed: 1.11] - click [required: Any, installed: 7.0] - click-default-group [required: Any, installed: 1.2.2] - click [required: Any, installed: 7.0] - tabulate [required: Any, installed: 0.8.5]

Python 3.6.8 ```

The default is to download everything. This can result in an enormous amount of data when you only really need 2 types of data for now:

• My Activity
• Location History

In addition unless you specify that "My Activity" is downloaded in JSON format the default is HTML. This then causes the

google-takeout-to-sqlite my-activity takeout.db takeout.zip

command to fail as it only contains html files not json files.

Thanks

Hi, Firstly, thanks for your work on datasette, it is a hugely useful tool!

I've been working on a fork [https://github.com/dazzag24/datasette-cluster-map] of datasette-cluster-map to allow the tileserver to be easily switched. Primarily because the tiles being served in the current version use localised text for labels and I'd like to have English used for these names instead.

It uses http://leaflet-extras.github.io/leaflet-providers/preview/ to allow you to simply set the tile provider using a call like so: let tiles = L.tileLayer.provider('Esri.WorldTopoMap'); instead of the current: let tiles = L.tileLayer('https://{s}.tile.openstreetmap.org/{z}/{x}/{y}.png', { maxZoom: 19, detectRetina: true, attribution: '© <a href="https://www.openstreetmap.org/copyright">OpenStreetMap</a> contributors' }), However I've got stuck in trying to work out how to pass the provider string to the plugin. In the documentation: https://datasette.readthedocs.io/en/stable/plugins.html you discuss configuration of plugins and use an example of passing in which latitude and longitude columns should be used. However I cannot seem to see anywhere in the current datasette-cluster-map code where these config params are passed in or used.

Can you please point me to an example or how to pass configuration from the metadata.json down into a plugin. Once I've over come this issue I was wondering if you would be interested in taking this change into your version?

Many thanks Darren