org.apache.hadoop.hive.ql.exec
Class ArchiveUtils

java.lang.Object
  org.apache.hadoop.hive.ql.exec.ArchiveUtils

public final class ArchiveUtils
extends Object

ArchiveUtils.

Nested Class Summary

static class	ArchiveUtils.HarPathHelper HarPathHelper helps to create har:/ URIs for locations inside of archive.
static class	ArchiveUtils.PartSpecInfo PartSpecInfo keeps fields and values extracted from partial partition info which is prefix of the full info.

Field Summary

static String

ARCHIVING_LEVEL

Constructor Summary

ArchiveUtils()

Method Summary

static String	addSlash(String s)
static URI	addSlash(URI u) Makes sure, that URI points to directory by adding slash to it.
static String	conflictingArchiveNameOrNull(Hive db, Table tbl, LinkedHashMap<String,String> partSpec) Determines if one can insert into partition(s), or there's a conflict with archive.
static int	getArchivingLevel(Partition p) Returns archiving level, which is how many fields were set in partial specification ARCHIVE was run for
static String	getPartialName(Partition p, int level) Get a prefix of the given parition's string representation.
static boolean	isArchived(Partition p) Determines whether a partition has been archived

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

ARCHIVING_LEVEL

public static String ARCHIVING_LEVEL

Constructor Detail

ArchiveUtils

public ArchiveUtils()

Method Detail

addSlash

public static String addSlash(String s)

addSlash

public static URI addSlash(URI u)
                    throws HiveException

Makes sure, that URI points to directory by adding slash to it. Useful in relativizing URIs.

Throws:

HiveException

isArchived

public static boolean isArchived(Partition p)

Determines whether a partition has been archived

Parameters:

p -

Returns:

is it archived?

getArchivingLevel

public static int getArchivingLevel(Partition p)
                             throws HiveException

Returns archiving level, which is how many fields were set in partial specification ARCHIVE was run for

Throws:

HiveException

getPartialName

public static String getPartialName(Partition p,
                                    int level)
                             throws HiveException

Get a prefix of the given parition's string representation. The sencond argument, level, is used for the prefix length. For example, partition (ds='2010-01-01', hr='00', min='00'), level 1 will reture 'ds=2010-01-01', and level 2 will return 'ds=2010-01-01/hr=00'.

Parameters:

p - partition object

level - level for prefix depth

Returns:

prefix of partition's string representation

Throws:

HiveException

conflictingArchiveNameOrNull

public static String conflictingArchiveNameOrNull(Hive db,
                                                  Table tbl,
                                                  LinkedHashMap<String,String> partSpec)
                                           throws HiveException

Determines if one can insert into partition(s), or there's a conflict with archive. It can be because partition is itself archived or it is to be created inside existing archive. The second case is when partition doesn't exist yet, but it would be inside of an archive if it existed. This one is quite tricky to check, we need to find at least one partition inside of the parent directory. If it is archived and archiving level tells that the archival was done of directory partition is in it means we cannot insert; otherwise we can. This method works both for full specifications and partial ones - in second case it checks if any partition that could possibly match such specification is inside archive.

Parameters:

db - - Hive object

tbl - - table where partition is

partSpec - - partition specification with possible nulls in case of dynamic partiton inserts

Returns:

null if partition can be inserted, string with colliding archive name when it can't

Throws:

HiveException