How to dynamically pass psql variable value at runtime? - postgresql

I am trying to create shortcuts for DBAs/admins in psql.
I have created some functions and queries, which give me information such as schema size in bytes, or a list of tables in order of storage used, etc. These will be used whenever an admin or DBA wants to get some info about the database.
I can run these queries naturally with select, or as a function like get_size(), but I want them to be accessible as shortcuts, similar to the native backslash commands (\dx, \dt, etc).
So I have used psql's \set feature to store queries/functions as variables, which I will put in the psqlrc file:
\set size 'select pg_size_pretty(my_size_function(''public''));'
Then when I type :size in psql I would get the size of the "public" schema.
What I want though, is to be able to dynamically pass a schema name, so I could run things like
:size public, :size schema2 etc.
I tried changing the \set to: \set size 'select pg_size_pretty(my_size_function(:schema));', but I can only call that by executing \set schema '''public''' first.
Since the whole point is to use these universally as shortcuts, having to manually run \set commands each time defeats the purpose.
In Oracle the would be colon a bind variable, which would be read at runtime.
How can I do this with psql?

I use these ways.
Postgres Functions (Inside Postgres CLI)
postgres=# CREATE OR REPLACE FUNCTION getSize(tableName varchar) RETURNS varchar LANGUAGE SQL as
postgres-# $$
postgres$# SELECT pg_size_pretty(pg_relation_size(tableName));
postgres$# $$;
CREATE FUNCTION
postgres=# select getSize('test');
getsize
------------
8192 bytes
(1 row)
From Shell (Outside Postgres CLI)
psqlgettablesize(schema, tablename) - Get table size. Pass all for schema argument to get for all schemas.
$ psqlgettablesize customer events
table_name | total_size | table_size | index_size
--------------------------+---------------------+-------------+---------------
test(complete database) | 19039892127 (18 GB) | |
--------- | --------- | --------- | ---------
customer.events | 24576 (24 kB) | 0 (0 bytes) | 24576 (24 kB)
(3 rows)
psqlgettablecount(schema, tablename) - Get count of rows in a table, in a schema
$ psqlgettablecount customer events
count
----------
51850000
(1 row)
psqlgetvacuumdetails(schema, tablename) - Get vacuum details of a table, in a schema
$ psqlgetvacuumdetails customer events
schemaname | relname | n_live_tup | n_dead_tup | last_analyze | analyze_count | last_autoanalyze | autoanalyze_count | last_vacuum | vacuum_count | last_autovacuum | autovacuum_count
------------+-----------+------------+------------+----------------------------+---------------+----------------------------+-------------------+---------------------------+--------------+-----------------+------------------
customer | events | 0 | 0 | 2019-12-02 18:25:04.887653 | 2 | 2019-11-27 18:49:19.002405 | 1 | 2019-11-29 13:11:15.92002 | 1 | | 0
(1 row)
psqltruncatetable(schema, tablename) - Truncate a table, in a schema, after authorization.
$ psqltruncatetable customer events
Are you sure to truncate table 'customer.events' (y/n)? y
Time: 4.944 ms
psqlsettings(category) - Get settings of Postgres
$ psqlsettings Autovacuum
name | setting | unit | category | short_desc | extra_desc | context | vartype | source | min_val | max_val | enumvals | boot_val | reset_val | sourcefile | sourceline | pending_restart
-------------------------------------+-----------+------+------------+-------------------------------------------------------------------------------------------+------------+------------+---------+---------+---------+------------+----------+-----------+-----------+------------+------------+-----------------
autovacuum | on | | Autovacuum | Starts the autovacuum subprocess. | | sighup | bool | default | | | | on | on | | | f
autovacuum_analyze_scale_factor | 0.1 | | Autovacuum | Number of tuple inserts, updates, or deletes prior to analyze as a fraction of reltuples. | | sighup | real | default | 0 | 100 | | 0.1 | 0.1 | | | f
psqlselectrows(schema, tablename) - Get rows from a table, in a schema
$ psqlselectrows customer events
id | name
----+------
1 | Clicked
2 | Page Loaded
(10 rows)
#Colors
B_BLACK='\033[1;30m'
B_RED='\033[1;31m'
B_GREEN='\033[1;32m'
B_YELLOW='\033[1;33m'
B_BLUE='\033[1;34m'
B_PURPLE='\033[1;35m'
B_CYAN='\033[1;36m'
B_WHITE='\033[1;37m'
RESET='\033[0m'
#Postgres Command With Params
psqlcommand="$POSTGRES_BIN/psql -U postgres test -q -c"
function psqlgettablesize()
{
[ -z "$1" ] && echo -e "${B_RED}Argument 1 missing. Schema name needed. ${B_YELLOW}(Pass 'all' to get details from all schema(s)${RESET}" ||
{
criteria="and table_schema = '$1'"
if [ "$1" == "all" ]; then
criteria=""
fi
if [ "$2" != "" ]; then
criteria+=" and table_name = '$2'"
fi
[ -z "$2" ] && echo -e "${B_YELLOW}Table name not given. ${B_GREEN}Showing size of all tables in $1 schema(s)${RESET}"
query="SELECT
concat(current_database(), '(complete database)') AS table_name, concat(pg_database_size(current_database()), ' (', pg_size_pretty(pg_database_size(current_database())), ')') AS total_size, '' AS table_size, '' AS index_size
UNION ALL SELECT '---------','---------','---------','---------'
UNION ALL SELECT
table_name,
concat(total_table_size, ' (', pg_size_pretty(total_table_size), ')'),
concat(table_size, ' (', pg_size_pretty(table_size), ')'),
concat(index_size, ' (', pg_size_pretty(index_size), ')')
FROM (
SELECT
concat(table_schema, '.', table_name) AS table_name,
pg_total_relation_size(concat(table_schema, '.', table_name)) AS total_table_size,
pg_relation_size(concat(table_schema, '.', table_name)) AS table_size,
pg_indexes_size(concat(table_schema, '.', table_name)) AS index_size
FROM information_schema.tables where table_schema !~ '^pg_' AND table_schema <> 'information_schema' $criteria ORDER BY total_table_size) AS sizes";
$psqlcommand "$query"
}
}
function psqlgettablecount()
{
[ -z "$1" ] && echo -e "${B_RED}Argument 1 missing: Need schema name${RESET}"
[ -z "$2" ] && echo -e "${B_RED}Argument 2 missing: Need table name${RESET}" ||
$psqlcommand "select count(*) from $1.$2;"
}
function psqlgetvacuumdetails()
{
[ -z "$1" ] && echo -e "${B_RED}Argument 1 missing: Need schema name${RESET}" ||
[ -z "$2" ] && echo -e "${B_RED}Argument 2 missing: Need table name${RESET}" ||
$psqlcommand "SELECT schemaname, relname, n_live_tup, n_dead_tup, last_analyze::timestamp, analyze_count, last_autoanalyze::timestamp, autoanalyze_count, last_vacuum::timestamp, vacuum_count, last_autovacuum::timestamp, autovacuum_count FROM pg_stat_user_tables where schemaname = '$1' and relname='$2';"
}
function psqltruncatetable()
{
[ -z "$1" ] && echo -e "${B_RED}Argument 1 missing: Need schema name${RESET}" ||
[ -z "$2" ] && echo -e "${B_RED}Argument 2 missing: Need table name${RESET}" ||
{
read -p "$(echo -e ${B_YELLOW}"Are you sure to truncate table '$1.$2' (y/n)? "${RESET})" choice
case "$choice" in
y|Y ) $psqlcommand "TRUNCATE $1.$2;";;
n|N ) echo -e "${B_GREEN}Table '$1.$2' not truncated${RESET}";;
* ) echo -e "${B_RED}Invalid option${RESET}";;
esac
}
}
function psqlsettings()
{
query="select * from pg_settings"
if [ "$1" != "" ]; then
query="$query where category like '%$1%'"
fi
query="$query ;"
$psqlcommand "$query"
if [ -z "$1" ]; then
echo -e "${B_YELLOW}Passing Category as first argument will filter the related settings.${RESET}"
fi
}
function psqlselectrows()
{
[ -z "$1" ] && echo -e "${B_RED}Argument 1 missing: Need schema name${RESET}" ||
[ -z "$2" ] && echo -e "${B_RED}Argument 2 missing: Need table name${RESET}" ||
$psqlcommand "SELECT * from $1.$2"
}

I couldn't figure out how to dynamically pass variables so I settled for a system of aliases and variables that work together to provide what I was looking for. So I can run :load_schema <schemaname> and then run commands interacting with the "loaded" schema. Loading a schema is simply aliasing a \set command. This way people who are unfamiliar with psql can just press : and tab autocomplete will show them all of my shortcuts.
All my code is here.
This isn't exactly what I was looking for so I won't accept it but it's close enough to post in case anyone else wants to do this.

Related

Using sql query in shell script

I have a shell script that uses output from a sql query and based on the value of one column sends out an alert. However i don't think it's capturing the value. although the value is not greater than 0 yet it still sends out an email.
Any idea where i am going wrong? Thanks.
............................................................................
#!/bin/sh
psql -d postgres -U postgres -c "select pid,application_name,pg_wal_lsn_diff(pg_current_wal_lsn(), sent_lsn) sending_lag,pg_wal_lsn_diff(sent_lsn,flush_lsn) receiving_lag,pg_wal_lsn_diff(flush_lsn, replay_lsn) replaying_lag,pg_wal_lsn_diff(pg_current_wal_lsn(), replay_lsn) total_lag from pg_stat_replication;"| while read total_lag;
do
echo $total_lag
lag1=$(echo $total_lag)
done
if [[ $lag1 -ge 0 ]]
then echo "Current replication lag is $lag1" |mail -s "WARNING!" abcd#mail.com
else
echo "No issue"
fi
............................................................................
this is the output of above query
pid | application_name | sending_lag | receiving_lag | replaying_lag | total_lag
-------+------------------+-------------+---------------+---------------+-----------
27823 | db123 | 0 | 0 | 0 | 0
27824 | db023 | 0 | 0 | 0 | 0

ERROR: user "xxx" cannot be dropped because permission dependency is found

I'm trying to drop a user from a Redshift cluster but receive the following error:
drop user "xxx";
ERROR: user "xxx" cannot be dropped because permission dependency is found
I've installed the admin views and revoked all privileges from all tables and schemas. I cannot find any reference to this specific error. It is also not included in this instructional: https://aws.amazon.com/premiumsupport/knowledge-center/redshift-user-cannot-be-dropped/
select ddl from admin.v_generate_user_grant_revoke_ddl where ddltype='revoke' and grantee='xxx' order by objseq, grantseq desc;
ddl
-----
(0 rows)
select ddl, grantor, grantee from admin.v_generate_user_grant_revoke_ddl where grantee='xxx' and ddltype='grant' and objtype <>'default acl' order by objseq,grantseq;
ddl | grantor | grantee
-----+---------+---------
(0 rows)
select * from pg_user where usename = 'xxx';
usename | usesysid | usecreatedb | usesuper | usecatupd | passwd | valuntil | useconfig
------------+----------+-------------+----------+-----------+----------+----------+-----------
xxx | 110 | f | f | f | xxx | |
(1 row)
select * from pg_default_acl where defacluser=110;
defacluser | defaclnamespace | defaclobjtype | defaclacl
------------+-----------------+---------------+-----------
(0 rows)
The user in not in any groups either. Any guidances is appreciated.
The user had not run queries for at least the last two weeks, he was not very active. His only access was through a Redshift/Excel ODBC setup. I did not check initially, but a day later there were no active sessions for him. I re-ran the drop user command and got the expected result. There must have been some lingering 'something'. For reference I ran this cmd to see who had active session: select * from stv_sessions;. My problem has been resolved by trying again the next day. https://docs.aws.amazon.com/redshift/latest/dg/r_STV_SESSIONS.html

postgres: Cannot use to_tsvector on a column named 'text'

I'm using postgres 11.5 and try to do the following:
update ccnc set fulltext_tokens = to_tsvector(title || '. ' || description || '. ' || text ) where fulltext_tokens is NULL;
Which results in
ERROR: column " text" does not exist
LINE 1: ...o_tsvector(title || '. ' || description || '. ' || text ) wh...
^
HINT: Perhaps you meant to reference the column "ccnc.text".
However, neither does using ccnc.text help:
felix=# update ccnc set fulltext_tokens = to_tsvector(title || '. ' || description || '. ' || ccnc.text) where fulltext_tokens is NULL;
ERROR: missing FROM-clause entry for table " ccnc"
LINE 1: ...o_tsvector(title || '. ' || description || '. ' || ccnc.text...
... nor is something odd with the respective column (no trailing spaces or the like):
felix=# \d+ ccnc
Table "public.ccnc"
Column | Type | Collation | Nullable | Default | Storage | Stats target | Description
-----------------+-----------------------------+-----------+----------+----------------------------------+----------+--------------+-------------
id | integer | | not null | nextval('ccnc_id_seq'::regclass) | plain | |
description | text | | | ''::text | extended | |
text | text | | | | extended | |
title | text | | | | extended | |
Edit:
Also quoting does not help, e.g.:
update ccnc set fulltext = title || '. ' || description || '. ' || "text";
ERROR: syntax error at or near ""text""
LINE 1: ... fulltext = title || '. ' || description || '. ' || "text";
I'd greatly appreciate any help on how to create that new column named fulltext_tokens. Thank you in advance :-)
Despite the very valuable comment by Richard, i.e., not using a column name such as "text" in the first place, one way to avoid the error is to use concat or concat_ws instead of the concat operator ||.
For instance:
update ccnc set fulltext = concat_ws('. ', title, description, text)

Dynamic Column name changes every year in where clause

I'm trying to automate a TSQL select statement on a website. Each year the column names change with the number values at the end of the name increasing by 1 so instead of manually updating the site I'm trying to figure out how to include the dynamic column name in the where clause.
The data looks something like this.
+------+------+------+------+------+
| FY18 | FY19 | FY20 | FY21 | FY22 |
+------+------+------+------+------+
| 1 | 0 | 1 | 0 | 1 |
| 1 | 1 | 0 | 1 | 0 |
| 1 | 0 | 0 | 1 | 0 |
+------+------+------+------+------+
Here is what I've come up with so far... The select statement looks something like this
Select distinct
POS
from TBL_Staff
where [' 'FY'+right(year(dateadd(month,3,getdate()))-1,2) '] = 1
What I'm trying to figure it out is if there is a way to dynamically generate the date and get SQl to recognize name+date as a column
Note: This is fake data so please let me know if something isn't clear
Any help on this is most appreciated.
You could use dynamic SQL
DECLARE #sql NVARCHAR(MAX) = 'Select distinct
POS
from TBL_Staff
where FY' + CAST(right(year(dateadd(month,3,getdate()))-1,2) AS VARCHAR) + ' = 1'
exec sp_executesql #sql
#sql is a variable that is used to construct a SQL statement dynamically. Then the sp_executesql procedure executes it. Beware of using dynamic sql when other alternatives exist. It's harder to read, can be difficult to maintain, and can be a security issue if taking input from users and you're not careful to sanitize input parameters.

How to preserve new line character while performing psql copy command

I have following content in my csv file(with 3 columns):
141413,"\"'/x=/></script></title><x><x/","Mountain View, CA\"'/x=/></script></title><x><x/"
148443,"CLICK LINK BELOW TO ENTER^^^^^^^^^^^^^^","model\
\
xxx lipsum as it is\
\
100 sometimes unknown\
\
travel evening market\
"
When I import above mentioned csv in mysql using following command, it treats the backslash() as new line; which is the expected behavior.
LOAD DATA INFILE '1.csv' INTO TABLE users FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '\"' LINES TERMINATED BY '\n';
MYSQL Output
But when I try to import to psql using copy command, it treats \ as a normal character.
copy users from '1.csv' WITH (FORMAT csv, DELIMITER ',', ENCODING 'utf8', NULL "\N", QUOTE E'\"', ESCAPE '\');
postgres Output
Try parsing these \ before importing the CSV file, e.g. using perl -pe or sed and the STDIN from psql:
$ cat 1.csv | perl -pe 's/\\\n/\n/g' | psql testdb -c "COPY users FROM STDIN WITH (FORMAT csv, DELIMITER ',', ENCODING 'utf8', NULL "\N", QUOTE E'\"', ESCAPE '\');"
This is how it looks like after the import:
testdb=# select * from users;
id | company | location
--------+-----------------------------------------+-------------------------------------------------
141413 | "'/x=/></script></title><x><x/ | Mountain View, CA"'/x=/></script></title><x><x/
148443 | CLICK LINK BELOW TO ENTER^^^^^^^^^^^^^^ | model +
| | +
| | xxx lipsum as it is +
| | +
| | 100 sometimes unknown +
| | +
| | travel evening market +
| |
(2 Zeilen)