I'm looking for an HBase shell command that will count the number of records in a specified column family.
I know I can run:
echo "scan 'table_name'" | hbase shell | grep column_family_name | wc -l
however this will run much slower than the standard counting command:
count 'table_name' , CACHE => 50000 (because the use of the CACHE=>50000)
and worse - it doesn't return the real number of records, but something like the total number of cells (if I'm not mistaken?) in the specified column family.
I need something of the sort:
count 'table_name' , CACHE => 50000 , {COLUMNS => 'column_family_name'}
Thanks in advance,
Michael
Here is Ruby code I have written when needed thing like you need. Appropriate comments are provided. It provides you with HBase
shell count_table
command. First parameter is table name and second is array of properties, the same as for scan
shell command.
Direct answer to your question is
count_table 'your.table', { COLUMNS => 'your.family' }
I also recommend to add cache, like for scan:
count_table 'your.table', { COLUMNS => 'your.family', CACHE => 10000 }
And here you go with sources:
# Argiments are the same as for scan command.
# Examples:
#
# count_table 'test.table', { COLUMNS => 'f:c1' }
# --- Counts f:c1 columsn in 'test_table'.
#
# count_table 'other.table', { COLUMNS => 'f' }
# --- Counts 'f' family rows in 'other.table'.
#
# count_table 'test.table', { CACHE => 1000 }
# --- Count rows with caching.
#
def count_table(tablename, args = {})
table = @shell.hbase_table(tablename)
# Run the scanner
scanner = table._get_scanner(args)
count = 0
iter = scanner.iterator
# Iterate results
while iter.hasNext
row = iter.next
count += 1
end
# Return the counter
return count
end