Can I automatically create a table in PostgreSQL from a csv file with headers?

Question

I'm running PostgreSQL 9.2.6 on OS X 10.6.8. I would like to import data from a CSV file with column headers into a database. I can do this with the COPY statement, but only if I first manually create a table with a column for each column in the CSV file. Is there any way to automatically create this table based on the headers in the CSV file?

Per this question I have tried

COPY test FROM '/path/to/test.csv' CSV HEADER;

But I just get this error:

ERROR: relation "test" does not exist

And if I first create a table with no columns:

CREATE TABLE test ();

I get:

ERROR: extra data after last expected column

I can't find anything in the PostgreSQL COPY documentation about automatically creating a table. Is there some other way to automatically create a table from a CSV file with headers?

Answer 1

You can't find anything in the COPY documentation, because COPY cannot create a table for you.
You need to do that before you can COPY to it.

Answer 2

There is a very good tool that imports tables into Postgres from a csv file. It is a command-line tool called pgfutter (with binaries for windows, linux, etc.). One of its big advantages is that it recognizes the attribute/column names as well.

The usage of the tool is simple. For example if you'd like to import myCSVfile.csv:

pgfutter --db "myDatabase" --port "5432" --user "postgres" --pw "mySecretPassword" csv myCSVfile.csv

This will create a table (called myCSVfile) with the column names taken from the csv file's header. Additionally the data types will be identified from the existing data.

A few notes: The command pgfutter varies depending on the binary you use, e.g. it could be pgfutter_windows_amd64.exe (rename it if you intend to use this command frequently). The above command has to be executed in a command line window (e.g. in Windows run cmd and ensure pgfutter is accessible). If you'd like to have a different table name add --table "myTable"; to select a particular database schema us --schema "mySchema". In case you are accessing an external database use --host "myHostDomain".

A more elaborate example of pgfutter to import myFile into myTable is this one:

pgfutter --host "localhost" --port "5432" --db "myDB" --schema "public" --table "myTable" --user "postgres" --pw "myPwd" csv myFile.csv

Most likely you will change a few data types (from text to numeric) after the import:

alter table myTable
  alter column myColumn type numeric
    using (trim(myColumn)::numeric)

Answer 3

There is a second approach, which I found here (from mmatt). Basically you call a function within Postgres (last argument specifies the number of columns).

select load_csv_file('myTable','C:/MyPath/MyFile.csv',24)

Here is mmatt's function code, which I had to modify slightly, because I am working on the public schema. (copy&paste into PgAdmin SQL Editor and run it to create the function)

CREATE OR REPLACE FUNCTION load_csv_file(
    target_table text,
    csv_path text,
    col_count integer)
  RETURNS void AS
$BODY$

declare

iter integer; -- dummy integer to iterate columns with
col text; -- variable to keep the column name at each iteration
col_first text; -- first column name, e.g., top left corner on a csv file or spreadsheet

begin
    set schema 'public';

    create table temp_table ();

    -- add just enough number of columns
    for iter in 1..col_count
    loop
        execute format('alter table temp_table add column col_%s text;', iter);
    end loop;

    -- copy the data from csv file
    execute format('copy temp_table from %L with delimiter '','' quote ''"'' csv ', csv_path);

    iter := 1;
    col_first := (select col_1 from temp_table limit 1);

    -- update the column names based on the first row which has the column names
    for col in execute format('select unnest(string_to_array(trim(temp_table::text, ''()''), '','')) from temp_table where col_1 = %L', col_first)
    loop
        execute format('alter table temp_table rename column col_%s to %s', iter, col);
        iter := iter + 1;
    end loop;

    -- delete the columns row
    execute format('delete from temp_table where %s = %L', col_first, col_first);

    -- change the temp table name to the name given as parameter, if not blank
    if length(target_table) > 0 then
        execute format('alter table temp_table rename to %I', target_table);
    end if;

end;

$BODY$
  LANGUAGE plpgsql VOLATILE
  COST 100;
ALTER FUNCTION load_csv_file(text, text, integer)
  OWNER TO postgres;

Note: There is a common issue with importing text files related to encoding. The csv file should be in UTF-8 format. However, sometimes this is not quite achieved by the programs, which try to do the encoding. I have overcome this issue by opening the file in Notepad++ and converting it to ANSI and back to UTF8.

Answer 4

I am using csvsql to generate the table layout (it will automatically guess the format):

head -n 20 table.csv | csvsql --no-constraints --tables table_name 

And then I use \COPY in psql. That's for me the fastest way to import CSV file.

You can also use sed with csvsql in order to get the desired datatype:

head -n 20 table.csv | csvsql --no-constraints --tables table_name  | sed 's/DECIMAL/NUMERIC/' | sed 's/VARCHAR/TEXT/'

Answer 5

Use sqlite as intermediate step.

Steps:

1. In the command prompt type: sqlite3
2. In the sqlite3 CLI type: .mode csv
3. .import my_csv.csv my_table
4. .output my_table_sql.sql
5. .dump my_table
6. Finally execute that sql in your Postgresql

Answer 6

I achieved it with this steps:

1. Convert the csv file to utf8

    iconv -f ISO-8859-1 -t UTF-8 file.txt -o file.csv

2. Use this python script to create the sql to create table and copy

#!/usr/bin/env python3
import csv, os
#pip install python-slugify
from slugify import slugify

origem = 'file.csv'
destino = 'file.sql'
arquivo = os.path.abspath(origem)

d = open(destino,'w')
with open(origem,'r') as f:

    header = f.readline().split(';')
    head_cells = []
    for cell in header:
        value = slugify(cell,separator="_")
        if value in head_cells:
            value = value+'_2'
        head_cells.append(value)
    #cabecalho = "{}\n".format(';'.join(campos))

    #print(cabecalho)
    fields= []
    for cell in head_cells:
        fields.append(" {} text".format(cell))
    table = origem.split('.')[0]
    sql = "create table {} ( \n {} \n);".format(origem.split('.')[0],",\n".join(fields))
    sql += "\n COPY {} FROM '{}' DELIMITER ';' CSV HEADER;".format(table,arquivo)

    print(sql)
    d.write(sql)

3.Run the script with

python3 importar.py

Optional: Edit the sql script to adjust the field types (all are text by default)

4. Run the sql script. Short for console

sudo -H -u postgres bash -c "psql mydatabase < file.sql" 

Answer 7

I haven't used it, but pgLoader (https://pgloader.io/) is recommended by the pgfutter developers (see answer above) for more complicated problems. It looks very capable.

Answer 8

For a single table, I did very simply, quickly and online through one of the many good converters that can be found on the web. Just google convert csv to sql online and choose one.