mneedham

import streamlit as st
import duckdb
from streamlit_searchbox import st_searchbox

atp_duck = duckdb.connect('atp.duck.db', read_only=True)

def search_players(search_term):
    query = '''
    SELECT DISTINCT winner_name AS player
    FROM matches

mneedham

import duckdb
import pandas as pd

con = duckdb.connect('atp-matches.db')

con.sql("INSTALL httpfs")
con.sql("LOAD httpfs")

csv_files = [
    f"https://raw.githubusercontent.com/JeffSackmann/tennis_atp/master/atp_matches_{year}.csv"

mneedham

CREATE OR REPLACE TABLE players
AS SELECT * FROM read_csv_auto('atp_players.csv', SAMPLE_SIZE=-1);

CREATE OR REPLACE TABLE rankings AS 
select * 
from 'atp_rankings_*.csv';

SELECT player_id, name_first, name_last
FROM players 
LIMIT 5;

mneedham

-- Fails because of weird date
CREATE TABLE players AS 
select *
from 'atp_players.csv';

-- all varchar
CREATE TABLE players1 AS 
select *
from read_csv_auto('atp_players.csv', ALL_VARCHAR=TRUE);

mneedham

SELECT count(*)
FROM 'data/*.parquet';

SELECT * 
FROM 'data/*.parquet' 
LIMIT 10;

DESCRIBE
SELECT * 
FROM 'data/yellow_tripdata_2011-07.parquet';

mneedham

# The NYC Taxis Dataset - https://www1.nyc.gov/site/tlc/about/tlc-trip-record-data.page

pip install parquet-cli

parq data/yellow_tripdata_2022-01.parquet

parq data/yellow_tripdata_2022-01.parquet --schema

parq data/yellow_tripdata_2022-01.parquet --head 10

mneedham

///usr/bin/env jbang "$0" "$@" ; exit $?
//DEPS org.apache.pinot:pinot-jdbc-client:0.8.0
//DEPS com.github.freva:ascii-table:1.2.0

// jbang Query.java "select * from customers limit 5"
// export JDBC_URL="jdbc:pinot://localhost:9000"

import static java.lang.System.*;
import java.sql.*;
import java.util.*;

mneedham

event: message

id: [{"topic":"eqiad.mediawiki.recentchange","partition":0,"timestamp":1635935095001},{"topic":"codfw.mediawiki.recentchange","partition":0,"offset":-1}]

data: {"$schema":"/mediawiki/recentchange/1.0.0","meta":{"uri":"https://pl.wikipedia.org/wiki/Kategoria:Infoboksy_%E2%80%93_brakuj%C4%85ce_parametry_%E2%80%93_Jednostka_administracyjna_infobox","request_id":"45eb5803-f73b-410c-a640-c76c5d86d91f","id":"5a7ee8b2-fead-4d54-b6fc-4bb433a08e51","dt":"2021-11-03T10:24:55Z","domain":"pl.wikipedia.org","stream":"mediawiki.recentchange","topic":"eqiad.mediawiki.recentchange","partition":0,"offset":3410637559},"id":119002871,"type":"categorize","namespace":14,"title":"Kategoria:Infoboksy – brakujące parametry – Jednostka administracyjna infobox","comment":"usunięto [[:Sukhaura]] z kategorii","timestamp":1635935095,"user":"MalarzBOT","bot":true,"server_url":"https://pl.wikipedia.org","server_name":"pl.wikipedia.org","server_script_path":"/w","wiki":"plwiki","parsedcomment":"usunięto <a href=\"/wiki/Sukh

mneedham

import streamlit as st
from pinotdb import connect
import pandas as pd
import time

st.title("GitHub Events")

broker_port = 8000
conn = connect(host='localhost', port=broker_port, path='/query/sql', scheme='http')

mneedham

import streamlit as st
from pinotdb import connect
import pandas as pd

st.title("GitHub Events")

broker_port = 8000
conn = connect(host='localhost', port=broker_port, path='/query/sql', scheme='http')

query = f"""