ccsvv

ccsvv (curses csv viewer) is a program that displays CSV (comma-separated values) files as a table. It supports:

• different field separators
• escaping with "
• multi line fields
• different line endings (win / mac / linux)
• unicode characters (wchar_t)
• header detection
• filtering
• sorting

See the manpage for the command line options and the commands that are supported by the running program.

Installation

Debian Package

The easiest way to install ccsvv is to the use the debian package attached to the latest release. Download the .deb file, check the md5 sum and install the package:

md5sum ccsvv_X.Y.Z_amd64.deb
sudo dpkg -i ccsvv_X.Y.Z_amd64.deb

Build from Sources

Ensure that the dependencies are fulfilled (see below). Download the sources zipfile for the latest release and compile ccsvv with:

make

make install

Dependencies

ccsvv uses ncursesw to create and display the tables. ncursesw is the wide character version of ncurses.

For the deb package or the precompiled executable the following list of dependencies is relevant for ubuntu 18.0.4:

• libc6 (>=2.27)
• libtinfo5 (>=6.1)
• libncursesw5 (>=6.1)

To successfully compile ccsvv requires a package with the library and debug informations and a developer package with the header files. For ubuntu 18.0.4 this is:

• libc6 (>=2.27)
• libtinfo5-dbg (>=6.1)
• libncursesw5-dbg (>=6.1)
• libncursesw5-dev (>=6.1)

If you compiled your own ncursesw5 lib, make sure that the LD_LIBRARY_PATH is properly set.

Installation from sources

The following document shows the installation of ncurses and ccsvv from their sources:

• Installation from sources

Example: /etc/passwd

ccsvv can be used to browse through all kinds of csv like files. The following example shows the /etc/passwd file, filtered by system.

ccsvv -d : /etc/passwd

Example: Derby DB

In this section we want to discuss how to use ccsvv to display the results of a database client.

We use Derby DB as our database, which has a client called ij. It accepts a property file, with basic configurations. In our case ij.properties:

ij.user=user
ij.password=password
ij.database=mydb
#ij.database=mydb;create=true
ij.protocol=jdbc:derby:
ij.showNoCountForSelect=true
ij.showNoConnectionsAtStart=true
ij.maximumDisplayWidth=256

For our example, we need to create an example database with a table. This can be done with a sql file create.sql with the following content:

create table mytab(idx int, name varchar(40));
insert into mytab values (1,'Derby DB'); 
insert into mytab values (2,'PostgreSQL'); 
insert into mytab values (3,'Maria DB'); 
insert into mytab values (4,'Mysql'); 

We can execute the statements with the following call. This requires the flag create=true to be set in the ij.properties file. Make sure that you removed the flag after you created the database, to prevent warnings.

ij -p ij.properties create.sql

Piping a select statement to the ij client returns the table data.

echo "select * from mytab;" | ij -p ij.properties 
ij version 10.14
ij> select * from mytab;
NUM        |ADDR                                    
----------------------------------------------------
1          |Derby DB                                
2          |PostgreSQL                              
3          |Maria DB                                
4          |Mysql                                   
ij>

This result is OK as long as the table is small.

To be able to use the output of ij for ccsvv, we have to remove the unnecessary lines. This can be done with a small shell script. The script is called with a sql statement and the first step is to copy that statement to a file. Using a file rather than piping the statement to the client gives better control over the promt of ij.

A call of grep removes the line, that separates the header from the table data, the ij prompt, the version information and lines that only consist of a single ;. This happens if you forget the ; character after your sql statement. The last line pipes the resulting data to ccsvv. The field separator is the pipe character |:

#!/bin/sh
set -u
echo "$@" > .derby.sql
result=$(ij -p ij.properties .derby.sql | grep -v -e '^-*$' -e '^ij>' -e '^ij version' -e '^;$')
echo "$result" | ./ccsvv -s -d '|' || echo "$result"

An example call of the script derby_client.sh with the result:

sh derby_client.sh "select * from mytab"

Further readings

There are some implementation details documented at the following pages:

• Header detection

• Implementation details

I use docker to build and verify installation packages and to ensure that the installation documentation is correct. A description can be found at:

• Docker support