Shantanu's Blog

Database Consultant

July 07, 2021

 

pandas case study 34

moving average for the variable length window


How do I calculate the moving average where the set of previous rows is not fixed? For e.g. this SQL query will calculate the average for the totalprice column for the last 1 month orders.

SELECT avg(totalprice) OVER (
    PARTITION BY custkey
    ORDER BY orderdate
    RANGE BETWEEN interval '1' month PRECEDING AND CURRENT ROW)
FROM orders

I have this dataframe:

from io  import StringIO
import pandas as pd

myst="""cust_1,2020-10-10,100
cust_2,2020-10-10,15
cust_1,2020-10-15,200
cust_1,2020-10-16,240
cust_2,2020-12-20,25
cust_1,2020-12-25,140
cust_2,2021-01-01,5
"""

u_cols=['custkey', 'orderdate', 'totalprice']

myf = StringIO(myst)
import pandas as pd
orders = pd.read_csv(StringIO(myst), sep=',', names = u_cols)
orders['orderdate'] = pd.to_datetime(orders['orderdate'])
df=df.sort_values(list(df.columns))


Answer:

orders['my_average'] = (orders.groupby('custkey')
                      .apply(lambda d: d.rolling('30D', on='orderdate')['totalprice'].mean())
                      .reset_index(level=0, drop=True)
                      .astype(int)
                   )

https://stackoverflow.com/questions/68268531/window-function-for-moving-average

Labels:


July 06, 2021

 

Quick test tables in Athena

If you quickly ant to test data using athena query, use with syntax as shown below:


WITH countries(country_code) AS (VALUES 'pol', 'CAN', 'USA')
SELECT upper(country_code) AS country_code
    FROM countries
 
In this example aggregate expression uses "OVER" function.
 
WITH students_results(student_id, result) AS (VALUES
    ('student_1', 17),
    ('student_2', 16),
    ('student_3', 18),
    ('student_4', 18),
    ('student_5', 10),
    ('student_6', 20),
    ('student_7', 16))
SELECT
    student_id,
    result,
    count(*) OVER (
        ORDER BY result
 ) AS close_better_scores_count
FROM students_results

Labels: ,


May 28, 2021

 

Athena and Unicode text

Athena supports unicode characters very well. For e.g. if the datafile looks like this...

"Root_word";"Word";"Primary";"Type";"Code";"Position";"Rule"
"अँटिबायोटिक","अँटिबायोटिक","अँटिबायोटिक","Primary","","",""
"अँटिबायोटिक","अँटिबायोटिकअंती","अँटिबायोटिक","Suffix","A","7293","001: 0 अंती ."
"अँटिबायोटिक","अँटिबायोटिकअर्थी","अँटिबायोटिक","Suffix","A","7293","002: 0 अर्थी ."
"अँटिबायोटिक","अँटिबायोटिकआतून","अँटिबायोटिक","Suffix","A","7293","003: 0 आतून ."
"अँटिबायोटिक","अँटिबायोटिकआतूनचा","अँटिबायोटिक","Suffix","A","7293","004: 0 आतूनचा ."
"अँटिबायोटिक","अँटिबायोटिकआतूनची","अँटिबायोटिक","Suffix","A","7293","005: 0 आतूनची ."
"अँटिबायोटिक","अँटिबायोटिकआतूनचे","अँटिबायोटिक","Suffix","A","7293","006: 0 आतूनचे ."
"अँटिबायोटिक","अँटिबायोटिकआतूनच्या","अँटिबायोटिक","Suffix","A","7293","007: 0 आतूनच्या ."
"अँटिबायोटिक","अँटिबायोटिकआतूनला","अँटिबायोटिक","Suffix","A","7293","008: 0 आतूनला ."

This create table statement is all I need...

create external table myptg (
root_word varchar(255),
derived_word varchar(255),
stemmed_word varchar(255),
type varchar(255),
code varchar(255),
position varchar(255),
rule varchar(255)
)
ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.OpenCSVSerde'
WITH SERDEPROPERTIES (
  'separatorChar' = '\;',
  'quoteChar' = '\"',
  'escapeChar' = '\\'
)
LOCATION 's3://ptg1/mc/'
TBLPROPERTIES ("skip.header.line.count"="1");

I can create a supporting table like this...

create external table gamabhana (derived_word varchar(255))
LOCATION 's3://ptg1/mc2/'
TBLPROPERTIES ("skip.header.line.count"="1");

A new table can be created using the syntax something like this...
 
create external table anoop (
serial_number int,
root_word varchar(255),
stem1_word varchar(255),
stem2_word varchar(255),
stem3_word varchar(255),
stem4_word varchar(255),
stem5_word varchar(255),
stem6_word varchar(255),
stem7_word varchar(255)
)
ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'
WITH SERDEPROPERTIES (
  'field.delim' = ',',
  'escape.delim' = '\\',
  'line.delim' = '\n'
)
LOCATION 's3://ptg1/mc3/'
TBLPROPERTIES ("skip.header.line.count"="1");


And then run a join statement like this...

create table gamabhana_match as
select a.derived_word, b.root_word, b.stemmed_word, b.type, b.code, b.position, b.rule, c.stem1_word, c.stem2_word, c.stem3_word, c.stem4_word, c.stem5_word, c.stem6_word, c.stem7_word
from gamabhana as a left join myptg as b
on b.derived_word = a.derived_word
left join anoop as c
on c.derived_word = a.derived_word

It will scan around 2 GB data (in this case) and the cost will be around 1 cent per query. This can also be done in MySQL. But importing data and building indexes is not easy. Unlike Athena, MySQL allows unlimited queries for free!

Athena is good for data that is important and accessed rarely.

Labels: , ,


May 18, 2021

 

Pandas case study 33

Let's assume we have 2 dataframes of english words.
The words may or may not be the same in both tables and the first dataframe has second column called "count". How do I merge these 2 dataframes and still get to know if a word is from first or second dataframe?

from io import StringIO
u_cols = ['word','count']
audit_trail = StringIO('''
test 1
testing 24
again 52
begin 6
''')

oscar = pd.read_csv(audit_trail, sep=" ", names = u_cols)
_____

from io import StringIO
u_cols = ['word' ]
audit_trail = StringIO('''
newer
age
computing
begin
''')

ptg = pd.read_csv(audit_trail, sep=" ", names = u_cols)

Use full outer join from merge method of pandas.
Do not forget to turn indicators on!

ndf=pd.merge(oscar,ptg,on="word", how="outer", indicator=True)

Labels:


May 03, 2021

 

Interactive Analysis of Sentence Embeddings

We can encode a sentence  into 768 dimensions using a pre-built model.
The data can be visualized using tensorflow projector.

https://projector.tensorflow.org/

Install the required python module:
#!pip install sentence-transformers

We will open a csv file having 2 columns. Text and label.
We will modify the dataframe to add outliers to the data.

import pandas as pd

df = pd.read_csv(
    "http://bit.ly/dataset-sst2", nrows=100, sep="\t", names=["text", "label"]
)

df["label"] = df["label"].replace({0: "negative", 1: "positive"})
df.loc[[10, 27, 54, 72, 91], "text"] = "askgkn askngk kagkasng"
df.to_csv("metadata.tsv", index=False, sep="\t")

Use the distilbert pre-built model to transform the text to return encoded values. Unlike other word vectors, this considers the entire sentence and return 768 dimensions irrespective of number of words per sentence.

from sentence_transformers import SentenceTransformer
sentence_bert_model = SentenceTransformer("distilbert-base-nli-stsb-mean-tokens")
e = sentence_bert_model.encode(df["text"])
embedding_df = pd.DataFrame(e)
embedding_df.to_csv("output.tsv", index=False, sep="\t", header=None)

The numpy array is converted to pandas dataframe and exported as csv.
The 2 files can be uploaded to tensorflow projector and you can easily find the outliers as shown in this blog post...

https://amitness.com/interactive-sentence-embeddings/

Labels: , ,


Archives

June 2001   July 2001   January 2003   May 2003   September 2003   October 2003   December 2003   January 2004   February 2004   March 2004   April 2004   May 2004   June 2004   July 2004   August 2004   September 2004   October 2004   November 2004   December 2004   January 2005   February 2005   March 2005   April 2005   May 2005   June 2005   July 2005   August 2005   September 2005   October 2005   November 2005   December 2005   January 2006   February 2006   March 2006   April 2006   May 2006   June 2006   July 2006   August 2006   September 2006   October 2006   November 2006   December 2006   January 2007   February 2007   March 2007   April 2007   June 2007   July 2007   August 2007   September 2007   October 2007   November 2007   December 2007   January 2008   February 2008   March 2008   April 2008   July 2008   August 2008   September 2008   October 2008   November 2008   December 2008   January 2009   February 2009   March 2009   April 2009   May 2009   June 2009   July 2009   August 2009   September 2009   October 2009   November 2009   December 2009   January 2010   February 2010   March 2010   April 2010   May 2010   June 2010   July 2010   August 2010   September 2010   October 2010   November 2010   December 2010   January 2011   February 2011   March 2011   April 2011   May 2011   June 2011   July 2011   August 2011   September 2011   October 2011   November 2011   December 2011   January 2012   February 2012   March 2012   April 2012   May 2012   June 2012   July 2012   August 2012   October 2012   November 2012   December 2012   January 2013   February 2013   March 2013   April 2013   May 2013   June 2013   July 2013   September 2013   October 2013   January 2014   March 2014   April 2014   May 2014   July 2014   August 2014   September 2014   October 2014   November 2014   December 2014   January 2015   February 2015   March 2015   April 2015   May 2015   June 2015   July 2015   August 2015   September 2015   January 2016   February 2016   March 2016   April 2016   May 2016   June 2016   July 2016   August 2016   September 2016   October 2016   November 2016   December 2016   January 2017   February 2017   April 2017   May 2017   June 2017   July 2017   August 2017   September 2017   October 2017   November 2017   December 2017   February 2018   March 2018   April 2018   May 2018   June 2018   July 2018   August 2018   September 2018   October 2018   November 2018   December 2018   January 2019   February 2019   March 2019   April 2019   May 2019   July 2019   August 2019   September 2019   October 2019   November 2019   December 2019   January 2020   February 2020   March 2020   April 2020   May 2020   July 2020   August 2020   September 2020   October 2020   December 2020   January 2021   April 2021   May 2021   July 2021  

This page is powered by Blogger. Isn't yours?