Generate Postgresql tables from CSV files
csvpg name.csv
csvpg parses a CSV file and prints a PostgreSQL syntax "create table"
statement into which the CSV can be imported.
By default it assumes that the first row of the CSV file is a header, and uses that to generate column names for the table. It then looks at the contents of each column to find an appropriate SQL type for it.
PostgreSQL supports NULL columns in CSV input, denoted by a zero length unquoted field. The CSV parser csvpg uses doesn't record whether a field is quoted or not, but if a zero length field is found then that SQL field will be generated as nullable.
There is untested support for generating enums for columns with a small number of values in it, disabled by default.
Usage of ./csvpg:
-columns string
comma-separated list of column names
-copy
include psql \copy command (default true)
-enums
consider enum as a type
-exclude string
exclude this comma-separated list of SQL types from consideration
-header
treat the first row of the file as a header (default true)
-notnull
check for not null (default true)
-sample int
consider only this many rows of the input
-snake
convert field names to snake_case (default true)
-table string
name of the generated table
With Go >= 1.11 installed, run "go build" from the cmd/csvpg subdirectory.