#Tesco Clone Script | Explore Tumblr posts and blogs

waiozseo · 5 years ago

Text

How TaskRabbit Clone have Impacted Business across the Globe

At around 2014, TaskRabbit hit the market hard and the demand for such applications across various other domains were on the rise. The use of the word Uber for X or TaskRabbit Clone Script became a very popular word across the online startup community.

What is a Taskrabbit Clone?

Taskrabbit clone is an online marketplace script that incorporates some ideas and features that made TaskRabbit successful.

Usually these apps house 2 different portals, one for the user and one for the service provider. They can be within the same app as in TaskRabbit or different apps like in uber.

The marketplace allows users to quickly order and use services like grocery, food, gardening, dog walking, etc, via a mobile app.

When a service request is submitted by the user, service providers in the vicinity get a notification on their app and accept to take up the service request.

The reason for such an online clone model to get the traction it did is because the model is simple to use, effective, and is very scalable. Also the concept of such on-demand service applications has disrupted the offline industry and has already created a huge ripple effect over the need to house an online business model. With safety and assurance over the service gotten, people started preferring such online models over offline businesses.

Industries which have made use of TaskRabbit clone modules

1. Food and grocery delivery

An essential commodity in everyday human life, food, and grocery delivery module has had a huge surge. There are umpteen online apps that cater to food delivery and grocery delivery across the globe.

Online Grocery delivery: The fact that we almost always forget the grocery run is the problem such on-demand grocery delivery apps. Every chain store right from Tesco, Costco, Asda, etc have their own e-commerce app. Apart from that you have certain standalone apps like Shipt, InstaCart, Peapod that have made inroads into the grocery business.

Food Delivery: Food delivery as such as had made it so much more convenient on days when you just want to stay home and do absolutely nothing. Takeaway industry in itself took a huge revolution when such apps came to prominence ever since 2016. Apps like uber eats, GrubHub, swiggy are testament to the industry.

2. Transportation and logistics

With taxi booking application already heavily influenced in the market. Uber, Lyft, Ola conquering hire taxi domains in most countries of the world. With this success, logistics and packers and mover also ventured into deploying such apps to develop their business.

Uber for Logistics: Logistics is one of the most in-demand domains in the world. The industry is filled with over 5.5 million businesses across the globe. Applications like Cargomatic, Lalamove trucky are all testament to how the industry is evolving.

Package delivery: Delivery of packages and couriers across cities has also taken up after the success of uber. Deploying real-time tracking of delivery and OTP verification has made the system very convenient and successful. Apps like dunzo, shyp etc are proof of how effective courier management is done via apps.

3. Housework

TaskRabbit created its demand on how well finding people to do odd jobs like cleaning, gardening, etc. Many applications have conceptualized TaskRabbit's idea and have ventured into their own domains.

Housekeeping: With the massive success of TaskRabbit, applications that help handymen and odd job services have taken over the globe. Bizzby, Handy, urban clap are examples of how deep these applications have penetrated the market. Be it gardening, window cleaning or maid services, TaskRabbit clones are your go-to option.

Care workers: Finding care workers for both nanny services and aged care have been made easy with such applications. You can see many applications that cater to babysitting across the globe these days.

Pet Care: Pet care industry has been on the rise and the need for dog-walking and pet care is now at a zenith. With GPS location of your dog and monthly plans applications like wag, Borrowmydoggy has taken the industry by the scruff forcing many others to look into such options.

Conclusion

Taskrabbit clone scripts or uber for X models are fast taking over industries that require a service provider and a consumer. It’s only a matter of time till your industry also deploys such applications to create a market revolution.

Beat your competitors to it by thinking about going digital with TaskRabbit clone scripts. To have a detailed consultation of how you can revolutionize your industry and stay ahead of your competitors, contact WAIOZ. WAIOZ has designed, crafted and developed a script with a huge amount of research done to assist you in standing out in your field.

#taskrabbit clone script

0 notes

goteso · 5 years ago

Text

Start Your Online Grocery Delivery Service With An App Like Tesco Groceries

Having a top supermarket shopping app like Tesco Groceries for your business store plays an important role in building a strong presence in the online grocery market. If you also want to gain a foothold as the best grocery delivery service and enhance your business management with software, consult with us at [email protected]. To boost your sales with a Tesco app clone, contact us today!

For more details, visit:

https://www.goteso.com/products/clone-scripts/best-online-grocery-shopping-and-delivery-app-like-tesco-groceries.html

#app like tesco groceries #tesco groceries app clone #grocery app development #grocery app #supermarket management software #grocery delivery service #best grocery delivery service #online grocery shopping and delivery #grocery shop app #online grocery market #grocery store sales app

0 notes

thirdyearfinalmajor · 6 years ago

Text

Going back to Shot 007

Starting to create my clean plate.

Looking through all my clean shots which I shoot, none of them works....In them, the camera was standing too far back on st mary’s street so the plotted tree’s are covering the information that I needed from the shops on the edges.

I can’t do much for now as its chucking it down with rain, but I have made plans with Paul to get some new clean plates tomorrow.

In the meantime, I want to see what I can create with paint out.

From this shot, I think the left-hand side would be the easiest to paint out.

Tried to paint it out the best I could. It hard with not much information to work with! I had to make sure to paint out anything that moves at a different speed within the parallax in this shot - like the tree and the lamp post

Rotoscope out the part i wantted to use

Projecting it back on

I had to go online and find a photo of the logo of the fantasy bar and projected

Colour Graded the sign

Started Rotoscoping the lamp post, the bench and the tree back in

What I have so far. Still got a lot of the tree to rotoscope out - but it's getting there. Once the tree is rotoscoped back in, the plain wall behind it, won't look so plain.

Bringing the table back. On my other paint out for this shot, the actually painted out the table, so I need to bring it back in. So I painted the top of the table out so I could blend it back into the clean plate.

Putting it on top of the wall clean plate

Adding the rotoscope on top.

One part is done! Next is the bit next door. This bit will hard because of the reflections.

Original

Paint Strokes

Clean

Roto

What it looks like through the scanline render

Roto on top

The backwall here is somewhere which i think i am going to struggle with so i am going to tackle this next.

I can see a bit of the floor here so I will use this to create the rest of the floor.

Pretty happy with that outcome. The reason why i have not continued to paint out the side is because the selected frame was not the best for the side window - only the window which i am focusing on. I will repaint that on a frame where i can see the bottom of the window.

The paint out process:

I haven’t really demonstrated how I painted the other one out so I am gonna show you on the other side of the plate (Tesco side) how I managed to clean it without having a good clean plate to put on top.

Original:

Here are all the paint strokes that I used to clean the plate. You can see that I completely covered the people. I started to paint out the tree and lampost. But then I released that I am not gonna replace the top of Tesco, I am gonna fade into that. (Means I don't have to roto out the tree!)

The first thing is first, I want to create that connection between the wall I am gonna paint out and the floor. There was a tiny bit that I could see so I just blended it over.

Blended it more up to create that bottom half of the wall.

Continued with the bottom wall.

I then went to paint out the first window. To be honest, I don’t know why, my eye was just attached to it, so I just started with that. Once again just cloning the rgb values down.

I have left the bottom out, as I need to look at how I am going to connect the window to the wall without looking too fake.

Painted out the door side, I tried to do the bottom of the window, but like i said it looks a bit fake.

Continued this process and cleaned up the bottom of the windows.

Here is the outcome of my clean plate for this bit.

Bringing back the plant pot - still, need to roto out the tree. the reason why I have not cleaned up the post is that I am going to put another plate on top of that.

I am going to come back to rotoscoping the tree out. I want to try and clean up the rest of this side first.

Next target: This will be hard because I don't have any floor details, I will have to make most of it up...

Using a tiny bit of floor I saw to spread it out

Bringing the floor up to cover the people’s legs

What I have done here is dragged the bottom of the floor from further back to create the floor for these buildings

Rotoscoping the A out from the sign

transforming it and merging it on top of the other sign

I have no idea what the test is meant to say so I am just painting it out.

Here is the outcome so far

Rotoscoping the bottom of the A to create the other side of it.

Rotoscoping this sign to replace on the other side

What I have done here is used my other clean plate from Tesco to duplicate one of the panels. I don’t have any reference for the cash point behind the lamp post so I am just replacing that.

Using roto paint to make it look more put together. + Painting out anything that has a different parallax.

Paint stokes

This section on my script.

The outcome. A very messy clean plate.

Rotoscoping the foreground elements back on top.

Looking at the rotoscope a hand goes past this part for like 4 frames

Using paint to frame by frame paint it out.

Done

Here is the result so far:

Overall, I am pretty happy with how it looks so far. I still need to rotoscope back the tree, but it is getting there! I also need to fix the shadow for that plant pot - but I will do that once all the people are out.

I think the next thing which I am going to tackle is the people right at the back in the middle.

For this back wall, I am going to break it up once. The dancer pretty much stayed in the middle of the road, so there is no need to make up the wall what is never seen.

Due to this, I have to choose a frame where the dancer is the furest to one side then the other. For the right, frame 1 would be the best choice to start on. Below is frame 1 and you can see how more windows are revealed and more pillars.

Also on this frame, I can see some details from the bottom of the windows. This is gonna help me cover the people and make it actually look like a window.

My paint lines

The outcome so far.

I stopped here because I don’t really have any more information to make the rest of the window. So I am just going to use what I have already created to make the rest of the window.

Masking out the side of the window which I think looks pretty good

Transforming it to the bottom.

I have done this to get that nice blend of colour between the frame and the wall.

of course, I cannot leave it like it is. So I used the roto paint to create that nice sharp corner.

Here are paint strokes.

Once again I am going to use this side to recreate the other side.

Transformed it over and rotated to create the slant

Think it is coming along nicely.

Fixing shadow

Now that window is done, i am going to mask it out and use it to create the other windows.

Of course, I don't want the windows to look the exact same so I am going to paint some of the shadows out to make it look different but still as if the sun is coming from the same angle. I have painted out some of the surrounding to make it blend better

Doing the same as before and using this window to make the next one.

I will be repeating the same process for this window.

Here is the result so far. Of course, I need to fix the edges which the sign was.

Masking out some of the walls from a different area to put over the mess.

Widen it to fix as well

Final paint to make everything blended together

Final result. I have left the signs in at some parts as I know I will not need them when the signs are back in, they will cover them. Also, they help me as a reference.

Coming together nicely - of course when the signs and pillars are back in, it will look more realistic.

Here is my first render:

youtube

From this render, I have found quite a few issues - but thankful they are all in the same area.

Looking at the right-hand side at Kaspers, the whole card slides, along with the middle pillar.

I can also see people move behind the tree as well as some of the tree moving and some staying still - I will need to rotoscope the whole tree out to move all of it.

Overcoming the issues:

Painting more of the tree out on the clean up of the back window.

Doing the same to this side aswell

Here is what i am left with. The reason why i have done this is so I can not see the people walking behind the tree / i can get the parralx movement better when the tree is rotscoped.

Render no 2

youtube

Again lots of issues. Mainly with the tree pot and the the bin at the back again

0 notes

globalmediacampaign · 5 years ago

Text

Building fast refresh capability in Amazon RDS for PostgreSQL

This post is a guest post by CDL. In their own words, CDL is a “leading insurtech company based in the UK, featured in the Financial Times Future 100 listing of high-growth UK companies making an impact on their industry. It has a strong track record in the insurance and financial services sectors, with its solutions powering some of the UK’s most profitable insurance retailers. Transacting over 7 million policies on its systems, CDL’s clients include Sainsbury’s Bank, Tesco Bank, Swinton Insurance, and Moneysupermarket.com.” This post describes how CDL developed fast refresh capability using materialized view logs in Amazon RDS for PostgreSQL. We detail what we built to track changes and offer an alternative to the complete refresh, which slashed the time needed from hours to seconds. We also share open-source software to enable fast refresh with the wider PostgreSQL community and outline the installation process involved. The challenge CDL handles millions of transactions each day. We were looking to migrate our business intelligence (BI) platform from Oracle to PostgreSQL. To make the switch viable, we needed the relational database to handle this volume of change and refresh in near-real time so that customers retain access to up-to-the-minute business intelligence. PostgreSQL only has the capability of a complete refresh. However, our service-level agreements require data to be refreshed every 15 minutes, and the volume of change transacted by CDL meant that the complete refresh process couldn’t handle the size of our materialized views (MVs) within this timescale. Our largest MV log is over 150 GB, which took hours to build through the complete refresh process, and there can be over 150 views per day, depending on the customer. We use the MVs on our BI solution to essentially de-normalize the data from our online transaction processing (OLTP) layer into logical data marts for our clients to consume and use to perform data analytics. Fast refresh capability was therefore an essential prerequisite for CDL when we switched from Oracle to PostgreSQL. Fast refresh vs. complete refresh Fast refresh uses materialized view logs on the underlying tables to keep track of changes, and only the changes since the last refresh are applied to the MV. In contrast, the complete refresh process refreshes all the data and could inevitably take hours for large datasets. The following diagram illustrates how we use MVs in our solution: we sync our data from the OLTP layer into an RDS instance using Attunity Replicate. We then build the MVs in the RDS instance from the OLTP source tables and refresh the jobs every 15 minutes using refresh schedules. Developing the process for PostgreSQL The fast refresh process was designed to be installed into its own schema that contains the functions needed to run the MV process, with three data dictionary tables and one role. The workflow for the MV log creation in the following diagram is as follows: Add the UUID column on the source table Create the MV log for the source table Add indexes on MV log table Create the trigger on the source table Insert the entry into the pgmview_logs table The workflow for the MV creation shown in the following diagram is as follows: Deconstruct the SQL statement Extract the compound MV tables Create the base table from MV SQL Add the rowid to the source table Add index to rowid column Insert into the pgmviews table Insert in the pgmviews_details table Complete refresh on the MV There are four main functions a user runs: mv$createMaterializedViewlog – Creates the MV logs mv$createMaterializedView – Creates the MV mv$removeMaterializedViewLog – Removes the MV logs mv$removeMaterializedView – Removes the MV Source code Download the source code from the GitHub repo. See the following code: git clone https://github.com/CDLSoftware/pg-mv-fast-refresh.git Installing the fast refresh functions The install of the fast refresh functions is designed to live in its own schema in the database that is specified via the MODULEOWNER parameter. To install the MV code, complete the following steps: Navigate to the folder where you downloaded the repo and edit the module_set_variable.sh file. This is where all the variables are stored for where you want to install the fast refresh functions. You don’t need the SOURCEUSERNAME/SOURCEPASSWORD and MVUSERNAME/MVPASSWORD parameters to install the fast refresh functions; you use them for the test harness setup. See the following code: cd pg-mv-fast-refresh vi module_set_variables.sh MODULEOWNER= - The module owner username MODULE_HOME= - The Module home path MODULEOWNERPASS= - Password for module owner PGRS_MVIEW HOSTNAME= - Hostname for database PORT= - port for database DBNAME= - Database Name PGUSERNAME= - DB username for the module installation run PGPASSWORD= - DB username password for the module installation run SOURCEUSERNAME= - DB username for the source tables for the MV SOURCEPASSWORD= - DB password for the source tables user MVUSERNAME= - DB username for the MV owner MVPASSWORD= - DB password for the MV owner LOG_FILE= - Path to logfile output location As an example of the parameter settings used in the test case, you have an RDS instance pg-tonytest.citnv5igau2v.eu-west-1.rds.amazonaws.com with a database testpoc and a master username dbamin. The install package installs the fast refresh functions under the schema testpoc. You have a source schema testpocsource. This is where the source data tables go for the test harness and a testpocmv, which is the schema where the MV is built. See the following code: ort MODULEOWNER=testpoc export MODULE_HOME=/var/lib/pgsql/pg-mv-fast-refresh export MODULEOWNERPASS=testpoc123 export HOSTNAME=pg-tonytest.citnv5igau2v.eu-west-1.rds.amazonaws.com export PORT=5432 export DBNAME=testpoc export PGUSERNAME=dbadmin export PGPASSWORD=testpoc123 export SOURCEUSERNAME=testpocsource export SOURCEPASSWORD=testpoc123 export MVUSERNAME=testpocmv export MVPASSWORD=testpoc123 export LOG_FILE=/tmp/fast_refresh_module_install_`date +%Y%m%d-%H%M`.log Change the permissions on the script runCreateFastRefreshModule.sh to execute and then run. See the following code: chmod 700 runCreateFastRefreshModule.sh ./runCreateFastRefreshModule.sh This should just take seconds to run. When it’s complete, check the log file in the location you set. For this post, the log file is in /tmp. The status is at the bottom. See the following example code: -bash-4.1$ cat fast_refresh_module_install_20190913-1358.log INFO: Set variables INFO: LOG_FILE parameter set to /tmp/fast_refresh_module_install_20190913-1358.log INFO: MODULEOWNER parameter set to testpoc INFO: PGUSERNAME parameter set to dbadmin INFO: HOSTNAME parameter set to pg-tonytest.citnv5igau2v.eu-west-1.rds.amazonaws.com INFO: PORT parameter set to 5432 INFO: DBNAME parameter set to strata INFO: MODULE_HOME parameter set to /var/lib/pgsql/pg-mv-fast-refresh INFO: Run testpoc schema build script INFO: Connect to postgres database strata via PSQL session …….. cut lines……….. GRANT INFO: Running Module Deployment Error Checks INFO: All Objects compiled successfully INFO: No Errors Found INFO: Completed Module Deployment Error Checks Seeing the functions in the database If you connect to the database on which you installed the package, you should see the functions under the MODULEOWNER. For this post, the functions are under testpoc. Installing the test harness The following use case demonstrates how to create a module view (MV). You create six tables and insert some data into the tables. All the following steps are in the create_test_harness.sh script and use the variables set in module_set_variables.sh. Create a schema for the test data. See the following code: psql --host=$HOSTNAME --port=$PORT --username=$PGUSERNAME --dbname=$DBNAME CREATE USER $SOURCEUSERNAME WITH LOGIN NOSUPERUSER NOCREATEDB NOCREATEROLE INHERIT NOREPLICATION CONNECTION LIMIT -1 PASSWORD '$SOURCEPASSWORD'; GRANT ALL PRIVILEGES ON DATABASE "$DBNAME" to $SOURCEUSERNAME; GRANT $SOURCEUSERNAME to $PGUSERNAME; CREATE SCHEMA $SOURCEUSERNAME AUTHORIZATION $SOURCEUSERNAME; GRANT ALL PRIVILEGES ON SCHEMA $SOURCEUSERNAME to $PGUSERNAME; GRANT $SOURCEUSERNAME to $MODULEOWNER; GRANT USAGE ON SCHEMA $SOURCEUSERNAME TO $MODULEOWNER; GRANT ALL ON SCHEMA $MODULEOWNER TO $SOURCEUSERNAME; Create a schema for the MV. See the following code: psql --host=$HOSTNAME --port=$PORT --username=$PGUSERNAME --dbname=$DBNAME CREATE USER $MVUSERNAME WITH LOGIN NOSUPERUSER NOCREATEDB NOCREATEROLE INHERIT NOREPLICATION CONNECTION LIMIT -1 PASSWORD '$MVPASSWORD'; GRANT $MVUSERNAME to $PGUSERNAME; CREATE SCHEMA IF NOT EXISTS $MVUSERNAME AUTHORIZATION $MVUSERNAME; GRANT ALL ON SCHEMA $MVUSERNAME TO $MVUSERNAME; GRANT SELECT ON ALL TABLES IN SCHEMA $SOURCEUSERNAME TO $MVUSERNAME; GRANT pgmv$_role TO $MVUSERNAME; ALTER DATABASE $DBNAME SET SEARCH_PATH=public,$MODULEOWNER,$MVUSERNAME,$SOURCEUSERNAME; GRANT $SOURCEUSERNAME TO $MODULEOWNER; GRANT USAGE ON SCHEMA $SOURCEUSERNAME TO $MODULEOWNER; GRANT ALL PRIVILEGES ON DATABASE strata TO $MODULEOWNER; GRANT ALL ON SCHEMA $MVUSERNAME TO $MODULEOWNER; GRANT USAGE ON SCHEMA $MVUSERNAME TO $MODULEOWNER; GRANT $MVUSERNAME TO $MODULEOWNER; GRANT $MODULEOWNER TO $MVUSERNAME; GRANT USAGE ON SCHEMA $MODULEOWNER TO $MVUSERNAME; GRANT ALL ON SCHEMA $SOURCEUSERNAME TO $MODULEOWNER; Create the test data as the sourceusername. See the following code: psql --host=$HOSTNAME --port=$PORT --username=$SOURCEUSERNAME --dbname=$DBNAME -- create t1 table CREATE TABLE $SOURCEUSERNAME.t1 ( id numeric NOT NULL, lookup_id numeric, code character varying(10) COLLATE pg_catalog."default", CONSTRAINT t1_pkey PRIMARY KEY (id) ); -- create t2 table CREATE TABLE $SOURCEUSERNAME.t2 ( id numeric NOT NULL, description character varying(100) COLLATE pg_catalog."default", metavals_id numeric, age integer NOT NULL, CONSTRAINT t2_pkey PRIMARY KEY (id) ); -- create t3 table CREATE TABLE $SOURCEUSERNAME.t3 ( lookup_id numeric NOT NULL, lookup_code character varying(10) COLLATE pg_catalog."default" NOT NULL, lookup_description character varying(50) COLLATE pg_catalog."default" NOT NULL, CONSTRAINT t3_pkey PRIMARY KEY (lookup_id) ); -- create t4 table CREATE TABLE $SOURCEUSERNAME.t4 ( metavals_id numeric NOT NULL, code character varying(10) COLLATE pg_catalog."default" NOT NULL, description character varying(30) COLLATE pg_catalog."default", CONSTRAINT t4_pkey PRIMARY KEY (metavals_id) ); -- create t5 table CREATE TABLE $SOURCEUSERNAME.t5 ( id numeric NOT NULL, rep_ind character varying(1) COLLATE pg_catalog."default", trans_id numeric, CONSTRAINT t5_pkey PRIMARY KEY (id) ); -- create t6 table CREATE TABLE $SOURCEUSERNAME.t6 ( trans_id numeric NOT NULL, payment_reference character varying(20) COLLATE pg_catalog."default" NOT NULL, CONSTRAINT t6_pkey PRIMARY KEY (trans_id) ); -- insert records into t1 table INSERT INTO $SOURCEUSERNAME.t1( id, lookup_id, code) VALUES (1, 10, 'hello'); INSERT INTO $SOURCEUSERNAME.t1( id, lookup_id, code) VALUES (2, 20, 'bye'); INSERT INTO $SOURCEUSERNAME.t1( id, lookup_id, code) VALUES (3, 30, 'cya'); INSERT INTO $SOURCEUSERNAME.t1( id, lookup_id, code) VALUES (4, 50, 'goodbye'); INSERT INTO $SOURCEUSERNAME.t1( id, lookup_id, code) VALUES (5, 50, 'hi'); INSERT INTO $SOURCEUSERNAME.t1( id, lookup_id, code) VALUES (6, 20, 'bye'); -- insert records into t2 table INSERT INTO $SOURCEUSERNAME.t2( id, description, metavals_id, age) VALUES (1, 'house', 100, 20); INSERT INTO $SOURCEUSERNAME.t2( id, description, metavals_id, age) VALUES (2, 'flat', 200, 35); INSERT INTO $SOURCEUSERNAME.t2( id, description, metavals_id, age) VALUES (3, 'bungalow', 300, 30); INSERT INTO $SOURCEUSERNAME.t2( id, description, metavals_id, age) VALUES (4, 'palace', 300, 30); INSERT INTO $SOURCEUSERNAME.t2( id, description, metavals_id, age) VALUES (5, 'office', 400, 50); -- insert records into t3 table INSERT INTO $SOURCEUSERNAME.t3( lookup_id, lookup_code, lookup_description) VALUES (10, 'ENG', 'ENGLAND'); INSERT INTO $SOURCEUSERNAME.t3( lookup_id, lookup_code, lookup_description) VALUES (20, 'WAL', 'WALES'); INSERT INTO $SOURCEUSERNAME.t3( lookup_id, lookup_code, lookup_description) VALUES (30, 'SCO', 'SCOTLAND'); INSERT INTO $SOURCEUSERNAME.t3( lookup_id, lookup_code, lookup_description) VALUES (40, 'IRE', 'IRELAND'); INSERT INTO $SOURCEUSERNAME.t3( lookup_id, lookup_code, lookup_description) VALUES (50, 'FRA', 'FRANCE'); -- insert records into t4 table INSERT INTO $SOURCEUSERNAME.t4( metavals_id, code, description) VALUES (100,'CHAIR','SMALL CHAIR'); INSERT INTO $SOURCEUSERNAME.t4( metavals_id, code, description) VALUES (200,'TABLE','SMALL TABLE'); INSERT INTO $SOURCEUSERNAME.t4( metavals_id, code, description) VALUES (300,'LIGHT','BRIGHT LIGHT'); INSERT INTO $SOURCEUSERNAME.t4( metavals_id, code, description) VALUES (400,'BED','KING SIZE BED'); INSERT INTO $SOURCEUSERNAME.t4( metavals_id, code, description) VALUES (500,'CUPBOARD','BEDSIDE CUPBOARD'); -- insert records into t5 table INSERT INTO $SOURCEUSERNAME.t5( id, rep_ind, trans_id) VALUES (1, 'Y', 1000); INSERT INTO $SOURCEUSERNAME.t5( id, rep_ind, trans_id) VALUES (2, 'Y', 2000); INSERT INTO $SOURCEUSERNAME.t5( id, rep_ind, trans_id) VALUES (3, 'N', 3000); INSERT INTO $SOURCEUSERNAME.t5( id, rep_ind, trans_id) VALUES (4, 'Y', 4000); INSERT INTO $SOURCEUSERNAME.t5( id, rep_ind, trans_id) VALUES (5, 'N', 5000); -- insert records into t6 table INSERT INTO $SOURCEUSERNAME.t6( trans_id, payment_reference) VALUES (1000, 'GZ-1000'); INSERT INTO $SOURCEUSERNAME.t6( trans_id, payment_reference) VALUES (2000, 'AZ-2000'); INSERT INTO $SOURCEUSERNAME.t6( trans_id, payment_reference) VALUES (3000, 'BZ-3000'); INSERT INTO $SOURCEUSERNAME.t6( trans_id, payment_reference) VALUES (4000, 'QZ-4000'); INSERT INTO $SOURCEUSERNAME.t6( trans_id, payment_reference) VALUES (5000, 'VZ-5000'); Create the MV logs. See the following code: psql --host=$HOSTNAME --port=$PORT --username=$SOURCEUSERNAME --dbname=$DBNAME DO $do$ DECLARE cResult CHAR(1) := NULL; BEGIN cResult := $MODULEOWNER.mv$createMaterializedViewlog( 't1','$SOURCEUSERNAME'); cResult := $MODULEOWNER.mv$createMaterializedViewlog( 't2','$SOURCEUSERNAME'); cResult := $MODULEOWNER.mv$createMaterializedViewlog( 't3','$SOURCEUSERNAME'); cResult := $MODULEOWNER.mv$createMaterializedViewlog( 't4','$SOURCEUSERNAME'); cResult := $MODULEOWNER.mv$createMaterializedViewlog( 't5','$SOURCEUSERNAME'); cResult := $MODULEOWNER.mv$createMaterializedViewlog( 't6','$SOURCEUSERNAME'); END $do$; Create the MV. See the following code: psql --host=$HOSTNAME --port=$PORT --username=$MVUSERNAME --dbname=$DBNAME DO $do$ DECLARE tStartTime TIMESTAMP := clock_timestamp(); cResult CHAR(1) := NULL; pSqlStatement TEXT; BEGIN pSqlStatement := ' SELECT t1.id t1_id, t1.lookup_id t1_lookup_id, t1.code t1_code, t2.id t2_id, t2.description t2_desc, t2.metavals_id t2_metavals_id, t2.age t2_age, t3.lookup_id t3_lookup_id, t3.lookup_code t3_lookup_code, t3.lookup_description t3_lookup_desc, t4.metavals_id t4_metavals_id, t4.code t4_code, t4.description t4_desc, t5.id t5_id, t5.rep_ind t5_rep_ind, t5.trans_id t5_trans_id, t6.trans_id t6_trans_id, t6.payment_reference t6_payment_ref FROM t1 INNER JOIN t2 ON t1.id = t2.id LEFT JOIN t3 ON t1.lookup_id = t3.lookup_id LEFT JOIN t4 ON t2.metavals_id = t4.metavals_id INNER JOIN t5 ON t1.id = t5.id LEFT JOIN t6 ON t5.trans_id = t6.trans_id'; cResult := mv$createMaterializedView ( pViewName => 'mv_fast_refresh_funct_test', pSelectStatement => pSqlStatement, pOwner => '$MVUSERNAME', pFastRefresh => TRUE ); RAISE NOTICE 'Complex Materialized View creation took % % %', clock_timestamp() - tStartTime, chr(10), chr(10); END $do$; Select data from the MV. See the following code: strata=> select t1_id, t1_lookup_id, t1_code, t2_id, t2_desc, t3_lookup_desc from mv_fast_refresh_funct_test order by t1_id; t1_id | t1_lookup_id | t1_code | t2_id | t2_desc | t3_lookup_desc -------+--------------+---------+-------+----------+---------------- 1 | 10 | hello | 1 | house | ENGLAND 2 | 20 | bye | 2 | flat | WALES 3 | 30 | cya | 3 | bungalow | SCOTLAND 4 | 50 | goodbye | 4 | palace | FRANCE 5 | 50 | hi | 5 | office | FRANCE (5 rows) Updating the test harness This is a very simple test to show how the fast refresh process works. For this post, you update a row in the base table t1 and see how that is reflected in the MV after a fast refresh. This test script is update_mv_test.sh. Update a row for table t1 with the code UPDATE testpocsource.t1 set code='yo' where code='hello': INFO: Changing a row for table t1 UPDATE testpocsource.t1 set code='yo' where code='hello' Press the enter key to do the update... UPDATE 1 Check that the MV the data isn’t changed (because it has not been refreshed). The following code shows that it is still set to hello: INFO: Check the output from the MV Select * from mv_fast_refresh_funct_test order by t1_id; Press the enter key to see the MV output... t1_id | t1_lookup_id | t1_code | t2_id | t2_desc | t3_lookup_desc -------+--------------+---------+-------+----------+---------------- 1 | 10 | hello | 1 | house | ENGLAND 2 | 20 | bye | 2 | flat | WALES 3 | 30 | cya | 3 | bungalow | SCOTLAND 4 | 50 | goodbye | 4 | palace | FRANCE 5 | 50 | hi | 5 | office | FRANCE (5 rows) Do a fast refresh on the MV and check it again. The following code shows that the row has changed: INFO: Let’s do a MV fast refresh to sync the MV with the change Press the enter key to do the MV refresh... NOTICE: Fast Snapshot Refresh took 00:00:00.055291 CONTEXT: PL/pgSQL function inline_code_block line 12 at RAISE DO INFO: Now Check the output from the MV the code for t1_id is now yo Press the enter key to see the MV output... t1_id | t1_lookup_id | t1_code | t2_id | t2_desc | t3_lookup_desc -------+--------------+---------+-------+----------+---------------- 1 | 10 | yo | 1 | house | ENGLAND 2 | 20 | bye | 2 | flat | WALES 3 | 30 | cya | 3 | bungalow | SCOTLAND 4 | 50 | goodbye | 4 | palace | FRANCE 5 | 50 | hi | 5 | office | FRANCE (5 rows) Change the row back to hello by updating the t1 table and refreshing the view again. See the following code: INFO: Now lets change the code back to hello UPDATE testpocsource.t1 set code='hello' where code='yo' Press the enter key to do the update... UPDATE 1 INFO: Lets do a MV fast refresh to sync the MV with the change Press the enter key to do the MV refresh... NOTICE: Fast Snapshot Refresh took 00:00:00.021894 CONTEXT: PL/pgSQL function inline_code_block line 12 at RAISE DO INFO: Now Check the output from the MV the code for t1_id its back to hello Press the enter key to see the MV output... t1_id | t1_lookup_id | t1_code | t2_id | t2_desc | t3_lookup_desc -------+--------------+---------+-------+----------+---------------- 1 | 10 | hello | 1 | house | ENGLAND 2 | 20 | bye | 2 | flat | WALES 3 | 30 | cya | 3 | bungalow | SCOTLAND 4 | 50 | goodbye | 4 | palace | FRANCE 5 | 50 | hi | 5 | office | FRANCE (5 rows) Removing the test harness To remove the test harness, run the drop_test_harness.sh script. This removes the data and MV schemas from the database. See the following code: -bash-4.2$ ./drop_test_harness.sh INFO: Drop Complete check logfile for status - /tmp/test_harness_drop_20191217-1300.log Removing the fast refresh functions To remove the fast refresh functions from the database, run the dropFastRefreshModule.sh script. This prompts you to enter Y to proceed with the removal of the objects. See the following code: -bash-4.2$ ./dropFastRefreshModule.sh Are you sure you want to remove the module schema - testpoc (y/n)?y yes selected the schemas - testpoc will be dropped Conclusion This post reviewed the fast refresh module and provided a simple example of getting an MV up and running. To enable our customers to access real-time business intelligence, we had to find a unique solution. We hope you enjoyed learning about how to develop fast refresh capabilities using MV logs with PostgreSQL. You can download the code from the GitHub repo. Stay tuned for part two of this series, where we’ll discuss the module and results we have seen in running this solution in the production environment. About the Author Tony Mullen is a Principal Data Engineer at CDL. He has over 18 years’ experience in databases; primarily Oracle, with a shift towards PostgreSQL over recent years. He heads up the migration of key applications from Oracle on-prem to Amazon RDS PostgreSQL, and has a keen interest in building the PostgreSQL community in Manchester, co-hosting the PostgreSQL Manchester meet-up. https://probdm.com/site/MjI5MDM

0 notes