CpSc 360

Lecture 10

Chapter 5 (continued)

· Reclaiming Space in Files

· Placement Strategies

· First-Fit: Go down the Avail Stack until a large enough slot is found

· Best-Fit:

· Keep list in Ascending order by size and use First-Fit.

· Residual after finding fit makes another smaller slot in list

· Search time can be significant for adds and deletes

· Worst-Fit: Keep list in Descending order by size and use First-Fit.

· Finding an open slot is fast!

· Residual after selection is large and won’t usually create much external fragmentation

· Conclusions:

· Fixed-length records do not require any special placement strategies.

· Variable length records in volatile files require careful placement.

· If space is lost due to internal fragmentation then consider first-fit or best-fit.

· If space is lost due to external fragmentation then consider worst-fit.

· Finding Things Quickly

· Sequential lookup is expensive

· Direct access with RRN or RBO require that we know a mapping from user data to RRN or RBO

· Binary searches are attractive for finding records by a key

· Problems with binary searches

· Require a sorted file

· Sorting is expensive – O(nlog₂n)

· External sorts require many seeks

· Internal sorts can help

· Requires log₂n + ½ accesses (average) to find a record – too many!

· Keeping a file sorted is very expensive

· Internal sorts only work on small files

· Keysorting

· Don’t sort the entire record but rather sort only the keys and addresses of the records [key, RRN] or [key, RBO] constructed during a sequential pass through the data file.

· Use RAM as much as possible to contain sort data

· Rearrange records in the file according to key sequence after sorting

For i:=1 to NumberOfRecords do

Seek Record (SortKey.RRN_i)

Read this record into a buffer in RAM

Write the contents of the buffer to output file

EndFor

· Problems with this technique

· Records must be read twice

· Many random seeks to disk on second pass through records

· If records have pointers to other records in them (i.e pinned records) then the rearrangement of the records may invalidate the pointers.

· Solution

· Don’t rearrange the records into sorted order

· Write the [key, RRN] back to a second (and smaller) file

· This file is actually an index to the original file

· A binary search can be performed on the index file much faster than the original

Chapter 6 – Indexing

· Advantages of indexes

· Enables one to impose an order on a file without actually rearranging the file

· Provide multiple access paths to a file (by multiple sort fields)

· Enables keyed access to variable-length record files

· The index is sorted

· The index will contain fixed length entries

· The index can be searched with a binary search

· The RBO data in the index entry can be used to access the target record

· The data file is called an entry-sequenced file

· Indexes are often small enough to be completely contained in memory

· Binary searches thus require no seeks and are extremely fast

· Algorithms required to use index files (assuming they will fit into RAM):

· Create the original empty index and data files

· Create empty data structures on disk

· Load the index file into memory before using it

· Simply read from disk into memory data structures

· Check for condition of the index and reconstruct index if possibly in damaged status

· Rewrite the index file from memory after using it

· Watch for power failures, etc.

· Post a status flag to note condition of index and protect against reading an out-of-date index

· Add records to the data file and index file

· Rearrangement of index entries into sorted order is likely

· Delete records from the data file

· Index entries must be deleted and index burped

· Update the records in the data file

· Updating non-key data

· Updating key data – must rearrange index entries to keep them sorted

· Retrieve records from the data file using the index file