如何从字符串中提取或测试最后一个字段？

Question

I am writing a bash script to test some data. I created a variable to hold the output of a SQL statement:

$ declare -p data_ck
declare -- data_ck="2019-02-17  TRUE    2019-02-10  23"

I presume my data_ck variable is really just a string rather than an array I can parse into fields. So, I next figured out this syntax: echo ${data_ck[@]:27:2}

which returns "23" in this case.

I could also use: echo "${data_ck[@]: -2:2}"

I want to determine whether my last element, 23, is equal to 30. I have tried different variations of this statement, without success:

If [ ${data_ck[@]:27:2} != 30 ]; then echo "missing dates"

which returns: -bash: syntax error near unexpected token `then'

I get the same result using this method:

If [ ${data_ck[@]:27:2} -ne 30 ]; then echo "missing dates" fi

I am only slightly familiar with scripting & do not understand what I am doing wrong. Can someone lend a hand? Thank you!

Answer 1

set -x
declare -- data_ck="2019-02-17  TRUE    2019-02-10  23"
last_field=${data_ck##*[[:space:]]}
[[ $last_field = 30 ]]

...properly emits:

+ [[ 23 = 30 ]]

...which tells us that ${data_ck##*[[:space:]]} successfully removed everything up to the last field, allowing the field's contents to be compared to 30 as per your specification.

It did this as a parameter expansion , matching and removing the longest-possible string that ends in a space. ( *[[:space:]] is a glob-style pattern matching any string ending with a space; ${var##pattern} expands to the contents of $var with the longest possible match for pattern removed from the beginning).

Answer 2

I can not see all variations of data_ck. It could be

data_ck=( 2019-02-17 TRUE 2019-02-10 23 )
data_ck=( 2019-02-17 FALSE 2019-02-10 23 )
data_ck=( 2019-02-17 TRUE 2019-02-10 301233 0 )
data_ck=( "2019-02-17 TRUE 2019-02-10 23" )

The value FALSE and the third example makes it impossible to use some fixed offset after converting the array to a string. You can extract the fourth field with

if [ "${data_ck[@]:3:1}" = "23" ]; then echo "23 found"; fi
# or shorter
if [ "${data_ck[3]}" = "23" ]; then echo "23 found"; fi

EDIT: When data_ck is a string, you still need to think what is the best way to het your field. Some possibilities:

echo "${data_ck}" | sed 's/.* //'    # Get everything after last space
echo "${data_ck}" | sed -r 's/.* ([^ ]+) */\1/' # get last field in string ending with spaces
echo "${data_ck}" | cut -d" " -f4    # Get 4th field
echo "${data_ck}" | awk '{print $4}' # 4th field where 2 spaces count as 1 delimiter

Answer 3

Assuming data_ck is a scalar variable, not an array, based on the result of delcare -p data_ck , you can just say:

data_ck="2019-02-17 TRUE 2019-02-10 23"
if [[ ${data_ck: -2:2} != 30 ]]; then
    echo "missing dates"
fi

The following still works but is mostly superfluous:

if [[ ${data_ck[@]: -2:2} != 30 ]]; then
    echo "missing dates"
fi

As @chepner points out, ${data_ck[@]} here is evaluated as ${data_ck[0]} and is equivalent to $data_ck in the context.