Abstract: In one embodiment, a method includes receiving an unstructured data record and parsing the unstructured data record to determine a characterization of the unstructured data record. The characterization includes determination of: a number of fields, a length for the fields, and an order of the fields as stored in the unstructured data record. The method also includes obtaining a first key pair that includes a primary key-name associated with a first key-value and hashing at least the first key-value to obtain a hash value. In addition, the method includes updating the unstructured data record to include the hash value as an indexing key thereby creating a modified data record. Also, the method includes storing the modified data record in a structured database. Moreover, the method includes updating an index of the structured database to include an entry for the modified data record, the entry including the hash value.

Abstract: In one embodiment, a method includes receiving a request to perform an update on a base data set in a record-oriented file system, wherein one or more affected records within the base data set are modified by the update. Affected alternate index (AIX) records corresponding to the affected records are located, and individual record-level locks are acquired on each affected AIX record. Each of the affected AIX records is updated according to the update after acquiring the individual record-level locks on each of the affected AIX records.

Abstract: In one embodiment, a method includes determining a data node corresponding to a data record to be inserted/updated in an index structure of a record-oriented file system, creating a lock on a corresponding data node, splitting the corresponding data node sequentially into two data nodes and storing the data record in one of the two data nodes in response to the corresponding data node not having free space sufficient to store/update the data record, storing a high key of the first of the two data nodes to a parent index node, splitting the parent index node sequentially into two index nodes in response to the parent index node not having free space sufficient to store/update the high key of the first of the two data nodes, and storing the high key of the first of the two data nodes in one of the two index nodes.

Abstract: In one embodiment, a method includes determining a data node corresponding to a data record to be inserted/updated in an index structure of a record-oriented file system, creating a lock on a corresponding data node, splitting the corresponding data node sequentially into two data nodes and storing the data record in one of the two data nodes in response to the corresponding data node not having free space sufficient to store/update the data record, storing a high key of the first of the two data nodes to a parent index node, splitting the parent index node sequentially into two index nodes in response to the parent index node not having free space sufficient to store/update the high key of the first of the two data nodes, and storing the high key of the first of the two data nodes in one of the two index nodes.

Abstract: In one embodiment, a process includes determining a data node for a data record to be inserted and/or updated in an index structure of a record-oriented file system. A lock on the corresponding data node is created, and the data record in the corresponding data node is stored and/or updated. However, when the corresponding data node does not have free space sufficient to store and/or update the data record, the corresponding data node is split sequentially into two data nodes. The new data record is stored in one of the two data nodes. The process continues by creating a lock on and updating a parent node in a sequence set which includes information about the corresponding data node and any parent index nodes above the parent index node which are affected by splitting the parent index node.

Abstract: In one embodiment, a method includes receiving a request to perform an update on a base data set in a record-oriented file system, wherein one or more affected records within the base data set are modified by the update. Affected alternate index (AIX) records corresponding to the affected records are located, and individual record-level locks are acquired on each affected AIX record. Each of the affected AIX records is updated according to the update after acquiring the individual record-level locks on each of the affected AIX records.

Abstract: In one embodiment, a method includes receiving a request to perform an update on a base data set in a record-oriented file system, locate all alternate index (AIX) records for the base data set, each AIX record including an alternate key field along with one or more relative byte address (RBA) values when the base data set is an entry sequenced data set (ESDS) or one or more primary keys when the base data set is a key sequenced data set (KSDS). Each RBA value/primary key corresponds to one record in the base data set. Individual record-level locks are acquired on each affected AIX record that includes alternate key field values and/or RBA values that are modified by the update. The affected AIX records may be updated after acquiring the individual record-level locks on each of the affected AIX records. Moreover, the affected AIX records may be updated in parallel with updating of unaffected AIX records, thereby greatly increasing performance.

Abstract: In one embodiment, a method includes receiving a request to perform an update on a base data set in a record-oriented file system, locate all alternate index (AIX) records for the base data set, each AIX record including an alternate key field along with one or more relative byte address (RBA) values when the base data set is an entry sequenced data set (ESDS) or one or more primary keys when the base data set is a key sequenced data set (KSDS). Each RBA value/primary key corresponds to one record in the base data set. Individual record-level locks are acquired on each affected AIX record that includes alternate key field values and/or RBA values that are modified by the update. The affected AIX records may be updated after acquiring the individual record-level locks on each of the affected AIX records. Moreover, the affected AIX records may be updated in parallel with updating of unaffected AIX records, thereby greatly increasing performance.

Abstract: In one embodiment, a process includes determining a data node for a data record to be inserted and/or updated in an index structure of a record-oriented file system. A lock on the corresponding data node is created, and the data record in the corresponding data node is stored and/or updated. However, when the corresponding data node does not have free space sufficient to store and/or update the data record, the corresponding data node is split sequentially into two data nodes. The new data record is stored in one of the two data nodes. The process continues by creating a lock on and updating a parent node in a sequence set which includes information about the corresponding data node and any parent index nodes above the parent index node which are affected by splitting the parent index node.