/*------------------------------------------------------------------------- * * hashjoin.h * internal structures for hash joins * * * Portions Copyright (c) 1996-2012, PostgreSQL Global Development Group * Portions Copyright (c) 1994, Regents of the University of California * * src/include/executor/hashjoin.h * *------------------------------------------------------------------------- */ #ifndef HASHJOIN_H #define HASHJOIN_H #include "nodes/execnodes.h" #include "storage/buffile.h" /* ---------------------------------------------------------------- * hash-join hash table structures * * Each active hashjoin has a HashJoinTable control block, which is * palloc'd in the executor's per-query context. All other storage needed * for the hashjoin is kept in private memory contexts, two for each hashjoin. * This makes it easy and fast to release the storage when we don't need it * anymore. (Exception: data associated with the temp files lives in the * per-query context too, since we always call buffile.c in that context.) * * The hashtable contexts are made children of the per-query context, ensuring * that they will be discarded at end of statement even if the join is * aborted early by an error. (Likewise, any temporary files we make will * be cleaned up by the virtual file manager in event of an error.) * * Storage that should live through the entire join is allocated from the * "hashCxt", while storage that is only wanted for the current batch is * allocated in the "batchCxt". By resetting the batchCxt at the end of * each batch, we free all the per-batch storage reliably and without tedium. * * During first scan of inner relation, we get its tuples from executor. * If nbatch > 1 then tuples that don't belong in first batch get saved * into inner-batch temp files. The same statements apply for the * first scan of the outer relation, except we write tuples to outer-batch * temp files. After finishing the first scan, we do the following for * each remaining batch: * 1. Read tuples from inner batch file, load into hash buckets. * 2. Read tuples from outer batch file, match to hash buckets and output. * * It is possible to increase nbatch on the fly if the in-memory hash table * gets too big. The hash-value-to-batch computation is arranged so that this * can only cause a tuple to go into a later batch than previously thought, * never into an earlier batch. When we increase nbatch, we rescan the hash * table and dump out any tuples that are now of a later batch to the correct * inner batch file. Subsequently, while reading either inner or outer batch * files, we might find tuples that no longer belong to the current batch; * if so, we just dump them out to the correct batch file. * ---------------------------------------------------------------- */ /* these are in nodes/execnodes.h: */ /* typedef struct HashJoinTupleData *HashJoinTuple; */ /* typedef struct HashJoinTableData *HashJoinTable; */ typedef struct HashJoinTupleData { struct HashJoinTupleData *next; /* link to next tuple in same bucket */ uint32 hashvalue; /* tuple's hash code */ /* Tuple data, in MinimalTuple format, follows on a MAXALIGN boundary */ } HashJoinTupleData; #define HJTUPLE_OVERHEAD MAXALIGN(sizeof(HashJoinTupleData)) #define HJTUPLE_MINTUPLE(hjtup) \ ((MinimalTuple) ((char *) (hjtup) + HJTUPLE_OVERHEAD)) /* * If the outer relation's distribution is sufficiently nonuniform, we attempt * to optimize the join by treating the hash values corresponding to the outer * relation's MCVs specially. Inner relation tuples matching these hash * values go into the "skew" hashtable instead of the main hashtable, and * outer relation tuples with these hash values are matched against that * table instead of the main one. Thus, tuples with these hash values are * effectively handled as part of the first batch and will never go to disk. * The skew hashtable is limited to SKEW_WORK_MEM_PERCENT of the total memory * allowed for the join; while building the hashtables, we decrease the number * of MCVs being specially treated if needed to stay under this limit. * * Note: you might wonder why we look at the outer relation stats for this, * rather than the inner. One reason is that the outer relation is typically * bigger, so we get more I/O savings by optimizing for its most common values. * Also, for similarly-sized relations, the planner prefers to put the more * uniformly distributed relation on the inside, so we're more likely to find * interesting skew in the outer relation. */ typedef struct HashSkewBucket { uint32 hashvalue; /* common hash value */ HashJoinTuple tuples; /* linked list of inner-relation tuples */ } HashSkewBucket; #define SKEW_BUCKET_OVERHEAD MAXALIGN(sizeof(HashSkewBucket)) #define INVALID_SKEW_BUCKET_NO (-1) #define SKEW_WORK_MEM_PERCENT 2 #define SKEW_MIN_OUTER_FRACTION 0.01 typedef struct HashJoinTableData { int nbuckets; /* # buckets in the in-memory hash table */ int log2_nbuckets; /* its log2 (nbuckets must be a power of 2) */ /* buckets[i] is head of list of tuples in i'th in-memory bucket */ struct HashJoinTupleData **buckets; /* buckets array is per-batch storage, as are all the tuples */ bool keepNulls; /* true to store unmatchable NULL tuples */ bool skewEnabled; /* are we using skew optimization? */ HashSkewBucket **skewBucket; /* hashtable of skew buckets */ int skewBucketLen; /* size of skewBucket array (a power of 2!) */ int nSkewBuckets; /* number of active skew buckets */ int *skewBucketNums; /* array indexes of active skew buckets */ int nbatch; /* number of batches */ int curbatch; /* current batch #; 0 during 1st pass */ int nbatch_original; /* nbatch when we started inner scan */ int nbatch_outstart; /* nbatch when we started outer scan */ bool growEnabled; /* flag to shut off nbatch increases */ double totalTuples; /* # tuples obtained from inner plan */ /* * These arrays are allocated for the life of the hash join, but only if * nbatch > 1. A file is opened only when we first write a tuple into it * (otherwise its pointer remains NULL). Note that the zero'th array * elements never get used, since we will process rather than dump out any * tuples of batch zero. */ BufFile **innerBatchFile; /* buffered virtual temp file per batch */ BufFile **outerBatchFile; /* buffered virtual temp file per batch */ /* * Info about the datatype-specific hash functions for the datatypes being * hashed. These are arrays of the same length as the number of hash join * clauses (hash keys). */ FmgrInfo *outer_hashfunctions; /* lookup data for hash functions */ FmgrInfo *inner_hashfunctions; /* lookup data for hash functions */ bool *hashStrict; /* is each hash join operator strict? */ Size spaceUsed; /* memory space currently used by tuples */ Size spaceAllowed; /* upper limit for space used */ Size spacePeak; /* peak space used */ Size spaceUsedSkew; /* skew hash table's current space usage */ Size spaceAllowedSkew; /* upper limit for skew hashtable */ MemoryContext hashCxt; /* context for whole-hash-join storage */ MemoryContext batchCxt; /* context for this-batch-only storage */ } HashJoinTableData; #endif /* HASHJOIN_H */

RE-INVENTING THE TELEVISION NEWS BUSINESS*

A revolution in video storytelling

Creating entirely new & cost-effective production methods

From the world leaders in video production training and the creators of Character Driven Storytelling™

*and every other business that uses video

WHAT WE DO

Over the past 35 years, we have designed, built or restructured some of the most powerful news and journalism companies in the world.

We replace the traditional TV news ‘crew’ with one highly trained journalist, working alone with nothing but an iPhone.

No more TV news ‘crews’, no editors and no field producers.

This is television news done the way newspaper journalism is done – one reporter with their electronic pad and pencil.

In doing this, we can cut the cost of production by as much as 75% while increasing ratings and audience engagement.

In the place of conventional TV news ‘packages’ – ie, reporter stand up, interview, b-roll, man on the street, we marry great journalism with Netflix and Hollywood storytelling.

It’s a combination that works.

We have taken most of our clients to #1 in their respective markets.

And it’s not just for news. Any company, any profit, any NGO and anyone else who is online needs to tell their story in compelling yet cost-effective video. We can teach you to do that. Either in person or virtually.

EXAMPLES OF WHAT WE CAN TEACH YOUR STAFF TO PRODUCE

ITAY HOD

Itay Hod, MMJ with KPIX/CBS in San Francisco, took the 5-Day Intensive Video Storytelling Bootcamp in 2018.

Because he works alone, with only an iPhone, he was able to embed himself with a homeless family.

Here’s the story he produced in a one-day turn.

KIET DO

Kiet Do, an MMJ with KPIX/CBS in San Francisco, took the 5-Day Intensive Video Storytelling Bootcamp in 2021.

Here is a story he produced, all on his own, with only an iPhone and in a one-day turn.

TAYLOR SCHAUB

Taylor Schaub, an MMJ with Spectrum News 1 in LA, took the 5-Day Intensive Video Storytelling Bootcamp in 2023.

Here is a story he turned in only one day, using only an iPhone. It was the first video story he ever did and it was nominated for an Emmy.

THE BOOTCAMP

How do we convert stations and whole networks to working in this way?

Since 1988, we have run intensive 5-Day Video Storytelling Bootcamps

We have done these all over the world.

These are hands-on bootcamps, and participants learn an entirely new way of creating TV news stories.

-We shoot at a 3:1 ratio or lower, so turnaround times are fast.

-We go directly from camera to timelilne and edit – no written scripts. We work in the medium of pictures and sound.

-We are entirely character-driven.

-We are driven by pictures and real events.

-We are focused almost entirely on ’the viewer experience’.

Since 1988, more than 70,000 journalists around the world have taken our bootcamps, either in person on virtualy.

Case Studies

CBS News

We have started to work with CBS News, bringing our ideas of character-driven storytelling to one of the most successful and biggest networks in the United States. Since beginning to work with them ratings have climbed and more importantly, audience engagement is through the ceiling.

Learn More

New York Times Television

We started New York Times Television in 1990 and it was the first paper to be brought into the world of TV. It quickly became one of the most successful non-fiction production companies in the United States. The series and documentaries we produced won many awards including multiple Emmys.

Learn More

The BBC

We have been working with the BBC since the year 2000 helping to convert their national news network to our visual storytelling technique. Most recently we have trained teams from their sports, documentaries, and comedy divisions to make character-driven stories using only smartphones.

Learn More

Spectrum News

For the past five years we have worked with Spectrum News to introduce and train their journalists on visual, character driven storytelling using smartphones helping to create a different kind of local news for their network of 24-Hour News Stations across the United States.

Learn More

The United Nations

In 2006, we were approached by the United Nations. Rather than rely on news outlets, it would be much easier to train the field operatives to produce their own stories. We spent two years working with the UN, training more than 100 of their staff in bootcamps in Geneva and Nairobi.

Learn More

The Newark Star-Ledger

We trained 50 print reporters at the paper to shoot and tell their own stories, in conjunction with their print work. We built a TV newsroom in their existing print newsroom – you could not ask for a better set and they began to live stream their stories in conjunction with their print work.

Learn More

Mcgraw Hill

We spent two years with McGraw Hill, training more than 150 of their staffers, making them completely video literate. McGraw/Hill media properties we transit included Business Week, Aviation Week, (what was the name of the architecture magazine), and JD Power and Associates.

Learn More

Voice of America

In 1990, we were approached by The Voice of America, the official broadcasting agency for the United States Government. When we met with VOA, they were only a short wave radio broadcaster, but working with them, we took them into television, launching VOA-TV.

Learn More

Oyster Yachts

British based Oyster Yachts makes some of the finest yachts in the world. Like every other company, they had to find a way to feed the never-ending video demands of social media – sites like Instagram and TikTok. We trained the Oyster staff to tell their own stories, using only iPhones.

Learn More

Scottish Environmental Protection Agency

We were approached by SEPA, the Scottish Environmental Protection Agency because they had to continually find a way to ‘feed the media beast’. The result was that SEPA was able to tell their own stories, whenever they wanted, and at almost no additional cost.

Learn More

Michael on Media

Michael Rosenblum has been writing about the media since 1988. His work and ideas have appeared in The Guardian, The Huffington Post, Ilkeston Life and many other publications.

He has been blogging regularly for the past 35 years on this subject. Having taught media studies at Columbia University, NYU and now the University of Oxford, he is considered an expert on this subject.

Continue reading this post or look back at previous posts.

Do You Have Questions About Learning Video Skills?

If you would like to know more about our courses contact us and one of our training advisors will be happy to call you.

Your IP : 216.73.216.134

RE-INVENTING THE TELEVISION NEWS BUSINESS*

A revolution in video storytelling

Creating entirely new & cost-effective production methods

From the world leaders in video production training and the creators of Character Driven Storytelling™

*and every other business that uses video

WHAT WE DO

Over the past 35 years, we have designed, built or restructured some of the most powerful news and journalism companies in the world.

We replace the traditional TV news ‘crew’ with one highly trained journalist, working alone with nothing but an iPhone.

No more TV news ‘crews’, no editors and no field producers.

This is television news done the way newspaper journalism is done – one reporter with their electronic pad and pencil.

In doing this, we can cut the cost of production by as much as 75% while increasing ratings and audience engagement.

In the place of conventional TV news ‘packages’ – ie, reporter stand up, interview, b-roll, man on the street, we marry great journalism with Netflix and Hollywood storytelling.

It’s a combination that works.

We have taken most of our clients to #1 in their respective markets.

And it’s not just for news. Any company, any profit, any NGO and anyone else who is online needs to tell their story in compelling yet cost-effective video. We can teach you to do that. Either in person or virtually.

EXAMPLES OF WHAT WE CAN TEACH YOUR STAFF TO PRODUCE

ITAY HOD

Itay Hod, MMJ with KPIX/CBS in San Francisco, took the 5-Day Intensive Video Storytelling Bootcamp in 2018.

Because he works alone, with only an iPhone, he was able to embed himself with a homeless family.

Here’s the story he produced in a one-day turn.

KIET DO

Kiet Do, an MMJ with KPIX/CBS in San Francisco, took the 5-Day Intensive Video Storytelling Bootcamp in 2021.

Here is a story he produced, all on his own, with only an iPhone and in a one-day turn.

TAYLOR SCHAUB

Taylor Schaub, an MMJ with Spectrum News 1 in LA, took the 5-Day Intensive Video Storytelling Bootcamp in 2023.

Here is a story he turned in only one day, using only an iPhone. It was the first video story he ever did and it was nominated for an Emmy.

THE BOOTCAMP

How do we convert stations and whole networks to working in this way?

Since 1988, we have run intensive 5-Day Video Storytelling Bootcamps

We have done these all over the world.

These are hands-on bootcamps, and participants learn an entirely new way of creating TV news stories.

-We shoot at a 3:1 ratio or lower, so turnaround times are fast.

-We go directly from camera to timelilne and edit – no written scripts. We work in the medium of pictures and sound.

-We are entirely character-driven.

-We are driven by pictures and real events.

-We are focused almost entirely on ’the viewer experience’.

Since 1988, more than 70,000 journalists around the world have taken our bootcamps, either in person on virtualy.

Case Studies

CBS News

New York Times Television

The BBC

Spectrum News

The United Nations

The Newark Star-Ledger

Mcgraw Hill

Voice of America

Oyster Yachts

Scottish Environmental Protection Agency

Michael on Media