lightest way to join 2 CSV (like an SQL LEFT JOIN) - windows

Question

I'm reling on community expertise to guide me in the best way about following topic.
In a professional context runing on windows without possibility to install MS-Office application I need to distribut to my team a way to join 2 CSV files and produce a 3rd CSV files as output. Exactly as if we run a SQL query like :
SELECT f1.*, f1.bar = f2.bar as baz
FROM CSVfile1 as f1
LEFT JOIN CSVfile2 as f2
ON f1.key = f2.key
Aims is currently reached with Excel + VBA but MS-office package will be removed and no more accessibly. A solution with MS-Acces is not envisageable because of the same reason.
The goal is to allow any body to actualise the 3rd CSV without any competence and specific installation on it Computer. So an approach with python or MS-SQL-Servr is not good also.
I was thinking to accomplish that with Powshell script but first. I'm not habit to use PowerShell but I can learn.
But before trying that I ask to the community if this is the best way ? Or if there is better solution ? (requirements: Windows OS (latest version), No MS-office, No specific install).
Thank you all.

Answer 1

PowerShell has no built-in join functionality (akin to SQL's[1]) as of v7.2, though adding a Join-Object cmdlet is being proposed in GitHub issue #14994; third-party solutions are available, via the PowerShell Gallery (e.g., JoinModule).
For now, if installing third-party tools isn't an option, you can roll your own solution with the following approach, which usesImport-Csv to load the CSV files, an auxiliary hashtable to find corresponding rows, and Add-Member to add columns (properties).
# Create sample CSV files.
$csv2 = #'
key,bar,quux
key1,bar1,quux1
key2,bar2,quux2
key3,bar3,quux3
'# > ./CSVFile1.csv
#'
key,bar
key1,bar1
key2,bar2a
'# > ./CSVFile2.csv
# Import the the 2nd file and load its rows
# (as objects with properties reflecting the columns)
# into a hashtable, keyed by the column 'key' values.
$hash = #{}
foreach ($row in Import-Csv ./CSVFile2.csv) {
$hash[$row.key] = $row
}
# Import the 1st file and process each row (object):
# Look for a matching object from the 2nd file and add
# a calculated column derived from both objects to the
# input object.
Import-Csv ./CSVFile1.csv | ForEach-Object {
$matching = $hash[$_.key]
$_ |
Add-Member -PassThru baz $(if ($matching) { [int] ($matching.bar -eq $_.bar) })
}
Pipe the last statement to Export-Csv to export the resulting objects to a CSV file.
(E.g.
... | Export-Csv -NoTypeInformation -Encoding utf8 Results.csv)
The above yields the following:
key bar quux baz
--- --- ---- ---
key1 bar1 quux1 1
key2 bar2 quux2 0
key3 bar3 quux3
[1] There is a -join operator, but its purpose is to join the elements of a single array to form a single string.

Answer 2

Here's an off the wall answer using the sqlite command-line shell (a single 900kb executable) and the same sql join command. https://sqlite.org/download.html Sqlite seems to have trouble with utf16 or "unicode" text files. Even Excel has more trouble importing a utf16 csv.
# making csv file with ">" (utf16) caused this error:
# CREATE TABLE csvfile1(...) failed: duplicate column name:
'key,bar,quux
key1,bar1,quux1
key2,bar2,quux2
key3,bar3,quux3' | set-content csvfile1.csv
'key,bar
key1,bar1
key2,bar2a' | set-content csvfile2.csv
'.mode csv
.import csvfile1.csv csvfile1
.import csvfile2.csv csvfile2
.headers on
SELECT f1.*, f1.bar = f2.bar as baz
FROM CSVfile1 as f1
LEFT JOIN CSVfile2 as f2
ON f1.key = f2.key' | .\sqlite3
# output
# key,bar,quux,baz
# key1,bar1,quux1,1
# key2,bar2,quux2,0
# key3,bar3,quux3,