Source data
I am working on an ELT project to load data from CSV files into PostgreSQL where I will transform it. The CSV files have many columns that are consistent across files, but also contain activity columns that are inconsistent with names like Date (05/19/2020), Type (05/19/2020), etc.
In the loading script I am merging all of the columns with dates in the column name into one jsonb column so I don't have to constantly add new columns to the raw data table.
The resulting jsonb column in the raw data table looks like this:
id
activity
12345678
{"Date (05/19/2020)": null, "Type (05/19/2020)": null, "Date (06/03/2020)": "06/01/2020", "Type (06/03/2020)": "E"}
98765432
{"Date (05/19/2020)": "05/18/2020", "Type (05/19/2020)": "B", "Date (10/23/2020)": "10/26/2020", "Type (10/23/2020)": "T"}
JSON to columns
Using the amazing create_jsonb_flat_view function from this post I can convert the jsonb to columns like this:
id
Date (05/19/2020)
Type (05/19/2020)
Date (06/03/2020)
Type (06/03/2020)
Type (10/23/2020
Date (10/23/2020)
Type (10/23/2020)
10629465
null
null
06/01/2020
E
98765432
05/18/2020
B
10/26/2020
T
Need to move part of column name to row
Now, this is where I'm stuck. I need to remove the portion of the column name that is the Activity Date (e.g. (05/19/2020)) and create a row for each id and ActivityDate with additional columns for Date and Type like this:
id
ActivityDate
Date
Type
12345678
05/19/2020
null
null
12345678
06/03/2020
06/01/2020
E
98765432
05/19/2020
05/18/2020
B
98765432
10/23/2020
10/26/2020
T
I followed your link to the create_jsonb_flat_view article yesterday and then forgot this question. While I thank you for pointing me there, I think that mentioning it worked against you.
A more conventional approach using regexp_replace() works here. I left the date values as strings, but you can convert them with to_date() if needed:
with parse as (
select id, e.k, e.v,
regexp_replace(e.k, '\s+\([0-9/]{10}\)', '') as k_no_date,
regexp_replace(e.k, '^.+([0-9/]{10}).+', '\1') as k_date_only
from rawinput
cross join lateral jsonb_each_text(activity) as e(k, v)
)
select id,
k_date_only as activity_date,
min(v) filter (where k_no_date = 'Date') as date,
min(v) filter (where k_no_date = 'Type') as type
from parse
group by id, k_date_only;
db<>fiddle here
#Mike-Organek's Answer works beautifully!
However, I was curious if the regexp_replace() calls might be slowing the query down a bit and it seemed I could get the same results using a simpler function.
Since Mike gave me a great example to start with I modified it to split on the space between Date and (05/19/2020).
For 20,000 rows, it went from taking an avg of 7 sec on my local machine to an avg of .9 sec.
Here is the resulting query:
with parse as (
select id, e.k, e.v,
split_part(e.k, ' ', 1) as k_no_date,
trim(split_part(e.k, ' ', 2),'()') as k_date_only
from rawinput
cross join lateral jsonb_each_text(activity) as e(k, v)
)
select id,
k_date_only as activity_date,
min(v) filter (where k_no_date = 'Date') as date,
min(v) filter (where k_no_date = 'Type') as type
from parse
group by id, k_date_only;
I am using golang, postgresql and masterminds/squirrel.
I am having two table customer with columns customer_id, customer_name, country and table customer_location with columns customer_id, region. There is an external map like the one given below to map country to region:
map:= {"belgium":"Europe","india":"Asia","argentina":"america"}
Can I use the value from the map above matching the key being the country value from customer table inside a join clause.
Something like this:
squirrel.StatementBuilder.PlaceholderFormat(squirrel.Dollar).
Select(`c.customer_name`).
From("customer c").
Join(fmt.Sprintf(`
customer_location cl ON c.customer_id = cl.customer_id AND cl.region = %v[%s]`,map,"c.country")).
I am creating a viewer for PostgreSQL. My SQL needs to sort on the type that is normal for that column. Take for example:
Table:
CREATE TABLE contacts (id serial primary key, name varchar)
SQL:
SELECT id::text FROM contacts ORDER BY id;
Gives:
1
10
100
2
Ok, so I change the SQL to:
SELECT id::text FROM contacts ORDER BY id::regtype;
Which reults in:
1
2
10
100
Nice! But now I try:
SELECT name::text FROM contacts ORDER BY name::regtype;
Which results in:
invalid type name "my first string"
Google is no help. Any ideas? Thanks
Repeat: the error is not my problem. My problem is that I need to convert each column to text, but order by the normal type for that column.
regtype is a object identifier type and there is no reason to use it when you are not referring to system objects (types in this case).
You should cast the column to integer in the first query:
SELECT id::text
FROM contacts
ORDER BY id::integer;
You can use qualified column names in the order by clause. This will work with any sortable type of column.
SELECT id::text
FROM contacts
ORDER BY contacts.id;
So, I found two ways to accomplish this. The first is the solution #klin provided by querying the table and then constructing my own query based on the data. An untested psycopg2 example:
c = conn.cursor()
c.execute("SELECT * FROM contacts LIMIT 1")
select_sql = "SELECT "
for row in c.description:
if row.name == "my_sort_column":
if row.type_code == 23:
sort_by_sql = row.name + "::integer "
else:
sort_by_sql = row.name + "::text "
c.execute("SELECT * FROM contacts " + sort_by_sql)
A more elegant way would be like this:
SELECT id::text AS _id, name::text AS _name AS n FROM contacts ORDER BY id
This uses aliases so that ORDER BY still picks up the original data. The last option is more readable if nothing else.
I have a 2 tables FIRST
id,rl_no,adm_date,fees
1,123456,14-11-10,100
2,987654,10-11-12,30
3,4343,14-11-17,20
and SECOND
id,rollno,fare,type
1,123456,20,bs
5,634452,1000,bs
3,123456,900,bs
4,123456,700,bs
My requirement is twofold,
1, i first need to get all columns from both tables with common rl_no. So i used:
SELECT a.ID,a.rl_no,a.adm_date,a.fees,b.rollno,b.fare,b.type FROM FIRST a
INNER JOIN
SECOND b ON a.rl_no = b.rollno
The output is like this:
id,rl_no,adm_date,fees,rollno,fare,type
1,123456,14-11-10,100,123456,20,bs
1,123456,10-11-12,100,123456,900,bs
1,123456,14-11-17,100,123456,700,bs
2,Next i wanted to get the sum(fare) of those rollno that were common between the 2 tables and also whose fare >= fees from FIRST table group by rollno and id.
My query is:
SELECT x.ID,x.rl_no,,x.adm_date,x.fees,x.rollno,x.type,sum(x.fare) as "fare" from (SELECT a.ID,a.rl_no,a.adm_date,a.fees,b.rollno,b.fare,b.type FROM FIRST a
INNER JOIN
SECOND b ON a.rl_no = b.rollno) x, FIRST y
WHERE x.rollno = y.rl_no AND x.fare >= y.fees AND x.type IS NOT NULL GROUP BY x.rollno,x.ID ;
But this is throwing in exceptions.
ORA-00979: not a GROUP BY expression
00979. 00000 - "not a GROUP BY expression"
The expected output will be like this:
id,rollno,adm_date,fare,type
1,123456,14-11-10,1620,bs
So could someone care to show an oracle newbie what i'm doing wrong here?
It looks like there's a couple different problems here;
Firstly, you're trying to group by an x.ID column which doesn't exist; it looks like you'll want to add ID to the selected columns in your sub-query.
Secondly, when aggregating with GROUP BY, all selected columns need to be either listed in the GROUP BY statement or aggregated. If you're grouping by rollno and ID, what do you want to have happen to all the extra values for adm_date, fees, and type? Are those always going to be the same for each distinct rollno and ID pair?
If so, simply add them to the GROUP BY statement, ie,
GROUP BY adm_date, fees, type, rollno, ID
If not, you'll need to work out exactly how you want to select which one to be output; If you've got output like your example (adding in an ID column here)
ID,adm_date,fees,rollno,fare,type
1,14-11-10,100,123456,20,bs
1,10-11-12,100,123456,900,bs
1,14-11-17,100,123456,700,bs
Call that result set 'a'. If I run;
SELECT a.ID, a.rollno, SUM(a.fare) as total_fare
FROM a
GROUP BY a.ID, a.rollno
Then the result will be a single row;
ID,rollno,total_fare
1,123456,1620
So, if you also select the adm_date, fees, and type columns, oracle has no idea what you mean to do with them. You're not using them for grouping, and you're not telling oracle how you want to pick which one to use.
You could do something like
SELECT a.ID,
FIRST(a.adm_date) as first_adm_date,
FIRST(a.fees) as first_fees,
a.rollno,
SUM(a.fare) as total_fare,
FIRST(a.type) as first_type
FROM a
GROUP BY a.ID, a.rollno
Which would give the result;
ID,first_adm_date,first_fees,rollno,total_fare,first_type
1,14-11-10,100,123456,1620,bs
I'm not sure if that's what you mean to do though.
I have a table with columns FIRSTNAME LASTNAME and I want to create a third column that combines those two columns into FIRSTNAME_LASTNAME but ALSO uses a special dictionary to convert some of the names. Say I just want to apply it to the FIRSTNAME, e.g.:
Albert -> Funnyguy, Kathleen -> Nerd, Megan -> Weirdo
So the new column for the "Albert Jones" row would be "Funnyguy_Jones".
Currently I do this in psycopg2 by reading in all the rows (in batches because the db is huge), using a python dictionary to convert and create the new column, then sending out the updates with UPDATE table SET newcol = tmp.newcol FROM (VALUES ...) etc. This is very slow because of reading it into python. Any tips?
EDIT: not all of the names have conversions (only like 10% of them do, for those I want to keep the original name)
If left join has a match COALESCE will choose t2.newName, other wise you will choose t1.firstName
SELECT t1.firstName,
t1.lastName,
COALESCE(t2.newName, t1.firstName) + '_' + t1.lastName as combinedName
FROM firstTable t1
LEFT JOIN newTable t2
ON t1.firstName = t2.firstName