0% found this document useful (0 votes)

139 views

Lazy Analysts Guide To Faster SQL

Write Faster SQL, Get More Done Introducing the Lazy Analyst's Guide to Faster SQL! We've taken what we've learned from thousands of hours working with SQL and put it into an easy-to-read guide. Learn tactics for pre-aggregating data, avoiding joins, minimizing table scans, and approximating results.

Uploaded by

Jon

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

139 views

Lazy Analysts Guide To Faster SQL

Uploaded by

Jon

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 18

From the SQL loving folks at:

selectdashboards.name,dashboards.name,log_counts.ct from dashboards log_countsfrom dashboards join (select distinc distinct_logs.dashboard_id, count(1) as ctfrom ( select distinct dashboard_id,user_id fromtime_on_site_ as
distinct_logs group by distinct_logs.daselectdashboards.name,dashboards.name,log_counts.ct from dashboards log_countsfrom dashboards join (select distinc distinct_logs.dashboard_id, count(1) as ctfrom ( select distinct
dashboard_id,user_id fromtime_on_site_ as distinct_logs group by distinct_logs.da nct_logs.daselectdashboards.name,dashboards.name,log_counts.ct from dashboards log_countsfrom dashboards join (select distinc
distinct_logs.dashboard_id, count(1) as ctfrom ( select distinct dashboard_id,user_id fromtime_on_site_ as distinct_logs group by distinct_logs.da nct_logs.daselectdashboards.name,dashboards.name,log_counts.ct from

Table of Contents
About Periscope and Authors .............................................. 1
Introduction ................................................................... 2
Pre-Aggregated Data ........................................................ 3
Reducing Your Data Set .................................................... 3
Materialized Views .......................................................... 4

Avoiding Joins ................................................................ 6

Generate_series ............................................................ 6
Window Functions .......................................................... 7

Avoiding Table Scans ........................................................ 9

Adding an Index ............................................................ 9
Moving Math in Where Clauses ........................................... 10

Approximations ..............................................................11
Hyperloglog ................................................................ 11
Sampling ................................................................... 14

Thanks for Reading! ........................................................ 16

About Periscope
Data analysts all over the world use Periscope to get and share insights fast. Our focus
on the people actually running the queries means you get a full featured SQL editor that
makes writing sophisticated queries quick and easy.
With our in-memory data caching system, customers realize an average of 150X
improvement in query speeds. Our intuitive sharing features allow you to quickly share
data throughout your organization. Just type SQL and get charts.
Periscope is built by a small team of hackers working out of a loft in San Francisco. We
love our customers, SQL, and that moment when a blip in the data makes you say, "wait
a minute"

The Authors
Jason Freidman

David Ganzhorn

After leaving Google to join Periscope,

Jason promptly began rebuilding
Periscope's backend services in Go. Before
Google, Jason was lord of backend

Ganz joined Periscope from Google,

where he ran dozens of A/B tests on the
search results page, generating billions
in new revenue. At Periscope he hacks

engineering at SF-based startup Tokbox.

on all levels of the stack.

Tom O'Neill

Harry Glaser

Periscope was started in Tom's apartment,

where he built the first version of the
product in a weekend. He leads Periscope's
engineering efforts, and holds the coveted
customer-facing bugfix record. (10
minutes!)

Harry handles Periscope's customer

support, sales, office hunting, whiskey
buying, and website copywriting. In his
spare time he still checks in a little code,
much to the annoyance of the other
engineers.

Page 1

Introduction
Were All About SQL at Periscope
At Periscope we spend a lot of time working with SQL. Our product is built around
creating charts through SQL queries. Every day we help customers write and debug
SQL. Were on a mission to create the worlds best SQL editor. Even our blog is all about
SQL!
Why We Made This Book
As part of all this SQL work, we are constantly thinking about how to make our queries
faster. One of the most common problems our customers encounter is slow SQL
queries. Since they come to us for help, weve built a lot of expertise around optimizing
SQL for faster queries.
Some of our most popular blog posts are about speeding up SQL and they consistently
get positive responses from the community. We figured people must be eager to learn
more about it, so we made this ebook hoping it'd be a useful tool for SQL analysts.
Whats in This Book
This book is divided into four sections: Pre-aggregated Data, Avoiding Joins, Avoiding
Table Scans and Approximations. Each section has tips that weve either covered on our
blog or written exclusively for this ebook. The tactics here vary from beginner to
advanced: Theres something for everyone!
Now go forth and make your queries faster!
Hugs and queries,
The Periscope Team

Page 2

Pre-Aggregating Data
Its common when running reports to need to combine data for a query from different
tables. Depending on where you do this in your process, you can be looking at a severely
slow query. At the simplest form an aggregate is a simple summary table that can be
derived by performing a group by SQL query.
Aggregations are usually precomputed, partially summarized data, stored in new
aggregated tables. The most ideal point to aggregate data for faster queries is
aggregating as early in the query as possible.
In this section well go over reducing a data set, grouping data, and materialized views.

Reducing Your Data Set

Count distinct is the bane of SQL analysts, so it was an
obvious choice for demonstrating aggregating and
reducing your data set. Let's start with a simple query we
run all the time: Which dashboards do most users visit?

In Periscope, this would give you a graph like this:

For starters, let's assume the handy indices on user_id

and dashboard_id are in place, and there are lots more
log lines than dashboards and users.
On just 10 million rows, this query takes 48 seconds. To
understand why, let's consult our handy SQL explain:

It's slow because the database is iterating over all the

logs and all the dashboards, then joining them, then
sorting them, all before getting down to real work of
grouping and aggregating.
Aggregate, Then Join
Anything after the group-and-aggregate is going to be a
lot cheaper because the data size is much smaller. Since
we don't need dashboards.name in the
group-and-aggregate, we can have the database do the
aggregation first, before the join:

Page 3

Pre-Aggregating Data

And now for the big reveal: This query takes 0.7 seconds!
That's a 28X increase over the previous query, and a 68X
increase over the original query.

This query runs in 20 seconds, a 2.4X improvement! Once

again, our trusty explain will show us why:

As always, data size and shape matters a lot. These

examples benefit a lot from a relatively low cardinality.
There are a small number of distinct (user_id,
dashboard_id) pairs compared to the total amount of
data. The more unique pairs there are the more data
rows that must be grouped and counted separately the
less opportunity there is for easy speedups.
Next time count distinct is taking all day, try a few
subqueries to lighten the load.

Materialized views
As promised, our group-and-aggregate comes before the
join. And, as a bonus, we can take advantage of the index
on the time_on_site_logs table.
First, Reduce The Data Set
We can do better. By doing the group-and-aggregate over
the whole logs table, we made our database process a lot
of data unnecessarily. Count distinct builds a hash set for
each group in this case, each dashboard_id to keep
track of which values have been seen in which buckets.

The best way to make your SQL queries run faster is to

have them do less work, and a great way to do less work
is to query a materialized view that's already done the
heavy lifting.
Materialized views are particularly nice for analytics
queries. For one thing, many queries do math on the
same basic atoms. For another, the data changes
infrequently (often as part of daily ETLs). Finally, those
ETL jobs provide a convenient home for view creation and
maintenance.
Redshift doesn't yet support materialized views out of the
box, but with a few extra lines in your import script (or a
tool like Periscope), creating and maintaining
materialized views as tables is a breeze.
Lifetime Daily ARPU (average revenue per user) is
common metric and often takes a long time to compute.
Let's speed it up with materialized views.

We've taken the inner count-distinct-and-group and

broken it up into two pieces. The inner piece computes
distinct (dashboard_id, user_id) pairs. The second piece
runs a simple, speedy group-and-count over them. As
always, the join is last.

Calculating Lifetime Daily ARPU

This common metric shows the changes in how much
money you're making per user over the lifetime of the
your product.

Page 4

Pre-Aggregating Data

For that we'll need a purchases table and a gameplays

table, and the lifetime accumulated values for each date.
Here's the SQL for calculating lifetime gameplays:

That's a monster query and it takes minutes to run on a

database with 2B gameplays and 3M purchases. That's
way to slow, especially if we want to quickly slice by
dimensions like what platform the game was played on.
Plus, similar lifetime metrics will need to recalculate the
same data over and over again!
Easy View Materialization on Redshift
Conveniently, we wrote our query in a format that makes
it obvious which parts can be extracted into materialized
views: lifetime_gameplays and lifetime_purchases.
We'll fake view materialization in Redshift by creating
tables, and Redshift makes it easy to create tables from
snippets of SQL:

The range join in the correlated subquery lets us

recalculate the distinct number of users for each date.
Here's the SQL for lifetime purchases in the same
format:

Do the same thing for lifetime_gameplays, and and

calculating Lifetime Daily ARPU now takes less than a
second to complete!

Now that the setup is done, we can calculate lifetime

daily ARPU:

Remember to drop and recreate these tables every time

you upload data to your Redshift cluster to keep them
fresh. Or create views in Periscope instead, and we'll
keep them up to date automatically!

Page 5

Avoiding Joins
Depending on how your database scheme is structured, youre likely to have data
required for common analysis queries in different tables. Joins are typically used to
combine this data. Theyre a powerful tool, yet they do have their downside.
Your database has to scan each table thats joined and figure how each row matches up.
This makes joins expensive when it comes to query performance. You can mitigate this
through smart join usage, but your queries would be even faster if you could completely
avoid joins.
In this section, well go over avoiding joins by using generate_series and window
functions.

Generate Series
Calculating Lifetime Metrics
Lifetime metrics are a great way to get a long-term
sense of the health of business. They're particularly
useful for seeing how healthy a segment is by comparing
to others over the long term.
For example, here's a fictional graph of lifetime game
players by platform:

We can see that iOS is our fastest-growing platform,

quickly surpassing the others in lifetime players despite
being launched months later!
However, metrics like these can be incredibly slow, since
they're typically calculated with distinct counts over
asymmetric joins. We'll show you how to calculate these
metrics in milliseconds instead of minutes.
Self-Joining
The easiest way to write the query is to join gameplays
onto itself. The first gameplays, which we name dates,
provides the dates that will go on our X axis. The second
gameplays, called plays, supplies every game play that
happend on or before each date.
The SQL is nice and simple:

Page 6

Avoiding Joins

This query has miserable performance: 37 minutes on a

10,000-row table!

Don't be fooled by the scans at the top. Those are just to

get the single min and max values for generate_series.

To see why, let's look at our explain:

Notice that instead of joining and nested-looping over

gameplays multiplied by gameplays, we are joining with
gameplays and a Postgres function, in this case generate_series. Since the series has only one row per date,
the result is a lot faster to loop over and sort.
But we can still do a lot better.

By asymmetrically joining gameplays to itself, we're

bogging down in an n^2 loop. Then we're sorting the
whole blown-out table!
Generate Series
Astute readers will notice that we're only using the first
gameplays table in the join for its dates. It conveniently
has the exact dates we need for our graph the dates of
every gameplay but we can select those separately and
then drop them into a generate_series.
The resulting SQL is still pretty simple:

Window Functions
Every user started playing on a particular platform on
one day. Once they started, they count forever. So let's
start with the first gameplay for each user on each
platform:

Now that we have that, we can aggregate it into how

many first gameplays there were on each platform on
each day:

This query runs in 26 seconds, achieving an 86X speedup

over the previous query!
Here's why:

We've included each user's first gameplay per platform in

a with clause, and then simply counted the number of
users for each date and platform.
Now that we have the data organized this way, we can
simply sum the values for each platform over time to get
the lifetime numbers! Because we need each date's sum
to include previous dates but not new dates, and we need
the sums to be separate by platform, we'll use a window
function.

Page 7

Avoiding Joins

Here's the full query:

As a result, this version runs in 28 milliseconds, a

whopping 928X speedup over the previous version, and an
unbelievable 79,000X speedup over the original self-join.
Window functions to the rescue!

The window function is worth a closer look:

Partition by platform means we want a separate sum for

each platform. order by dt specifies the order of the rows,
in this case by date. That's important because rows
between unbounded preceding and current row specifies
that each row's sum will include all previous rows (e.g.
all previous dates), but no future rows. Thus it becomes a
rolling sum, which is exactly what we want.
With the full query in hand, let's look at the explain:

No joins at all! And when we're sorting, it's only over the
relatively small daily_first_gameplays with-clause, and
the final aggregated result.

Page 8

Avoiding Table Scans

Most SQL queries that are written for analysis without performance in mind will cause
the database engine to look at every row in the table. After all, if you want to aggregate
rows that meet certain conditions, what better way than to see whether each row meets
those conditions!
Unfortunately, as datasets get bigger, looking at every single row just won't work in a
reasonable amount of time. This is where it helps to give the query planner the tools it
needs to only look at a subset of the data. This section will give you a few building
blocks that you can use in queries that are scanning more rows than they should.

The Power of Indices

Most conversations about SQL performance begin with
indices, and rightly so. They possess a remarkable
combination of versatility, simplicity and power. Before
proceeding with many of the more sophisticated
techniques in this book, it's worth understanding what
they are and how they work.
Table Scans
Unless otherwise specified, the table is stored in the
order the rows were inserted. Unfortunately, this means
that queries that only use a small subset of the data for
example, queries about today's data must still scan
through the entire table looking for today's rows.
As an example, let's look at a real-world query. Let's look
at today's activity by hour, as measured by the pings table
we log events to:

Placeholder

This query takes 74.7 seconds. We can see why by asking

Postgres to explain the query:

The operation on the left-hand side of the arrow is a full

scan of the pings table! Because the table is not sorted
by created_at, we still have to look at every row in the
table to decide whether the ping happened today or not.
Indexing the data
Let's improve things by adding an index on created_at:

This will create a pointer to every row, and sort those

points by the columns we're indexing on in this case,
created_at. The exact structure used to store those
pointers depends on your index type, but binary trees are
very popular for their versatility and log(n) search times.

Page 9

Avoiding Table Scans

Now that we have an index, let's take a look at how the

query works:

Ignoring The Index

As usual, running explain will tell us why it's slow: It
turns out our database is ignoring the index and doing a
sequential scan!

Instead of scanning the whole table, we just search the

index! To top it off, now that it can ignore the pings that
didn't happen today, the query runs in 1.7s a 44X
improvement!
Caveats
Creating a lot of indices on a table will slow down inserts
on that table. This is because every index has to be
updated every time data is inserted. So proceed with
care! A good strategy is to take a look at your most
common queries, and create indices designed to
maximize efficiency on those queries.

Moving Math in Where

Clauses
A Deceptively Simple Query
Front and center on our Periscope dashboard is a
question we ask all the time: How much usage has there
been today?

The problem is that we're doing math on our indexed

column, created_at. This causes the database to look at
each row in the table, compute created_at - interval '8
hour' for that row, and compare the result to date(now() interval '8 hour').
Moving Math To The RHS
Our goal is to compute one value in advance, and let the
database search the created_at index for that value. To
do that, we can just move all the math to the
right-hand-side of the comparison:

This query runs in a blazing 85ms. With a simple change,

we've achieved a 3,000X speedup!
As promised, with a single value computed in advance,
the database can search its index:

To answer it, we've written a seemingly simple query:

Notice the "- interval '8 hour'" operations in the where

clause. Times are stored in UTC, but we want "today" to
mean today PST.

Of course, sometimes the math can't be refactored quite

so easily. But when writing queries with restricts on
indexed columns, avoid math on the indexed column. You
can get some pretty massive speedups.

As the time_on_site_logs table has grown to 22M rows,

even with an index on created_at, this query has slown
way down to an average of 267 seconds!

Page 10

Approximations
For truly large data sets, approximations can be one of the most powerful performance
tools in your toolset. In certain contexts, an answer that is accurate to +/- 5% is just as
useful for making a decision and multiple orders of magnitude faster to get.
More than most techniques, this one comes with a few caveats. The first is making sure
the customers of your data understand the limitations. Error bars are a useful
visualization that are well-understood by data consumers.
The second caveat is to understand the assumptions your technique makes about the
distribution of your data. For example, if your data isn't normally distributed, sampling is
going to produce incorrect results.
For more details about various technique and their uses, read on!

Hyperloglog
We'll optimize a very simple query, which calculates the
daily distinct sessions for 5,000,000 gameplays
(~150,000/day):

The original query takes 162.2s. The HyperLogLog

version is 5.1x faster (31.5s) with a 3.7% error, and uses a
small fraction of the RAM.
Why HyperLogLog?
Databases often implement count(distinct) in two ways:
When there are few distinct elements, the database
makes a hashset in RAM and then counts the keys.

When there there are too many elements to fit in RAM,

the database writes them to disk, sorts the file, and then
counts the number of element groups. The second case
writing the intermediate data to disk is very slow.
Probabilistic counters are designed to use as little RAM
as possible, making them ideal for large data sets that
would otherwise page to disk.
The HyperLogLog Probabilistic Counter is an algorithm
for determining the approximate number of distinct
elements in a set using minimal RAM. Distincting a set
that has 10 million unique 100-character strings can take
over gigabyte of RAM using a hash table, while
HyperLogLog uses less than a megabyte (the "log log" in
HyperLogLog refers to its space efficiency). Since the
probabilistic counter can stay entirely in RAM during the
process, it's much faster than any alternative that has to
write to disk and usually faster than alternatives using a
lot more RAM.

Page 11

Approximations

Hashing

Bucketing

The core of the HyperLogLog algorithm relies on one

simple property of uniform hashes: The probability of the
position of the leftmost set bit in a random hash is 1/2n,
where n is the position. We call the position of the
leftmost set bit the most significant bit, or MSB.

The maximum MSB for our elements is capable of

crudely estimating the number of distinct elements in the
set. If the maximum MSB we've seen is 3, given the
probabilities above we'd expect around 8 (i.e. 23) distinct
elements in our set. Of course, this is a terrible estimate
to make as there are many ways to skew the data.

Here are some hash patterns and the positions of their

MSBs:

We'll use the MSB position soon, so here it is in SQL:

Hashtext is an undocumented hashing function in

postgres. It hashes strings to 32-bit numbers. We could
use md5 and convert it from a hex string to an integer,
but this is faster.

The HyperLogLog algorithm divides the data into

evenly-sized buckets and takes the harmonic mean of the
maximum MSBs of those buckets. The harmonic mean is
better here since it discounts outliers, reducing the bias
in our count.
Using more buckets reduces the error in the distinct
count calculation, at the expense of time and space. The
function for determining the number of buckets needed
given a desired error is:

We'll aim for a +/- 5% error, so plugging in 0.05 for the

error rate gives us 512 buckets. Here's the SQL for
grouping MSBs by date and bucket:

We use ~(1 << 31) to clear the leftmost bit of the hashed
number. Postgres uses that bit to determine if the
number is positive or negative, and we only want to deal
with positive numbers when taking the logarithm.
The floor(log(2,...)) does the heavy lifting: The integer part
of base-2 logarithm tells us the position (from the right)
of the MSB. Subtracting that from 31 gives us the
position of the MSB from the left, starting at 1.
With that line we've got our MSB per-hash of the
session_id field!

The hashtext(...) & (512 - 1) gives us the rightmost 9 bits,

511 in binary is 111111111), and we're using that for the
bucket number.
The bucket_hash line uses a min inside the logarithm
instead of something like this max(31 - floor(log(...))) so
that we can compute the logarithm once - greatly
speeding up the calculation.
Now we've got 512 rows for each date - one for each
bucket - and the maximum MSB for the hashes that fell
into that bucket. In future examples we'll call this select
bucketed_data.

Page 12

Approximations

Counting
It's time to put together the buckets and the MSBs. The
paper linked above has a lengthy discussion on the
derivation of this function, so we'll only recreate the
result here. The new variables are m (the number of
buckets, 512 in our case) and M (the list of buckets
indexed by j, the rows of SQL in our case). The denominator of of this equation is the harmonic mean mentioned
earlier:

And the SQL for that looks like this:

Now putting it all together:

In SQL, it looks like this:

We add in (512 - count(1)) to account for missing rows. If

no hashes fell into a bucket it won't be present in the
SQL, but by adding 1 per missing row to the result of the
sum we achieve the same effect.
The num_zero_buckets is pulled out for the next step
where we account for sparse data.
Almost there! We have distinct counts that will be right
most of the time - now we need to correct for the
extremes. In future examples we'll call this select
counted_data.

And that's the HyperLogLog probabilistic counter in pure

SQL!

Correcting

The HyperLogLog algorithm really shines when you're in

an environment where you can count distinct in parallel.
The results of the bucketed_data step can be combined
from multiple nodes into one superset of data, greatly
reducing the cross-node overhead usually required when
counting distinct elements across a cluster of nodes. You
can also preserve the results of the bucketed_data step
for later, making it nearly free to update the distinct count
of a set on the fly!

The results above work great when most of the buckets

have data. When a lot of the buckets are zeros (missing
rows), then the counts get a heavy bias. To correct for
that we apply the formula below only when the estimate
is likely biased, with this equation:

Bonus: Parallelizing

Page 13

Approximations

Sampling
Sampling is an incredibly powerful tool to speed up
analyses at scale. While it's not appropriate for all
datasets or all analyses, when it works, it really works. At
Periscope, we've realized several orders of magnitude in
speedups on large datasets with judicious use of
sampling.
However, when sampling from databases, it's easy to lose
all your speedups by using inefficient methods to select
the sample itself. In this post we'll show you how to
select random samples in fractions of a second.

Query Faster by Sorting Only a Subset of the Table

The most obvious way to speed this up is to filter down
the dataset before doing the expensive sort.
We'll select a larger sample than we need and then limit
it, because we might get randomly fewer than the
expected number of rows in the subset. We also need to
randomly sort afterward to avoid biasing towards earlier
rows in the table.
Here's our new query:

The Obvious, Correct, Slow solution

Let's say we want to send a coupon to a random hundred
users as an experiment. Quick, to the database!

This baby runs in 7.97s: twice as fast!

The naive approach sorts the entire table randomly and

selects N results. It's slow, but it's simple and it works
even when there are gaps in the primary keys.
Selecting a random row:

This Query is Taking Forever!

On a Postgres database with 20M rows in the users table,
this query takes 17.51 seconds! To find out why, let's
return to our trusty explain:

The database is sorting the entire table before selecting

our 100 rows! This is an O(n log n) operation, which can
easily take minutes or longer on a 100M+ row table. Even
on medium-sized tables, a full table sort is unacceptably
slow in a production environment.

This is pretty good, but we can do better. You'll notice

we're still scanning the table, albeit after the restriction.
Our next step will be to avoid scans of any kind.
Generate random indices in the ID range
Ideally we wouldn't use any scans at all, and rely entirely
on index lookups. If we have an upper bound on table
size, we can generate random numbers in the ID range
and then lookup the rows with those IDs.

This puppy runs in 0.064s, a 273X speedup over the naive

query!

selectdashboards.name,dashboards.name,log_counts.ct from dashboards log_countsfrom dashboards join (select distinc distinct_logs.dashboard_id, count(1) as ctfrom ( select distinct dashboard_id,user_id fromtime_on_site_
as distinct_logs group by distinct_logs.daselectdashboards.name,dashboards.name,log_counts.ct from dashboards log_countsfrom dashboards join (select distinc distinct_logs.dashboard_id, count(1) as ctfrom ( select distinct
dashboard_id,user_id fromtime_on_site_ as distinct_logs group by distinct_logs.da nct_logs.daselectdashboards.name,dashboards.name,log_counts.ct from dashboards log_countsfrom dashboards join (select distinc
distinct_logs.dashboard_id, count(1) as ctfrom ( select distinct dashboard_id,user_id fromtime_on_site_ as distinct_logs group by distinct_logs.da nct_logs.daselectdashboards.name,dashboards.name,log_counts.ct from

Page 14

Approximations

Counting the table itself takes almost 8 seconds, so we'll

just pick a constant beyond the end of the ID range,
sample a few extra numbers to be sure we don't lose any,
and then select the 100 we actually want.
Bonus: Random sampling with replacement
Imagine you want to flip a coin a hundred times. If you flip
a heads, you need to be able to flip another heads. This is
called sampling with replacement. All of our previous
methods couldn't return a single row twice, but the last
method was close: If we remove the inner group by id,
then the selected ids can be duplicated:

Sampling is an incredibly powerful tool for speeding up

statistical analyses at scale, but only if the mechanism
for getting the sample doesn't take too long. Next time
you need to do it, generate random numbers first, then
select those records. Or, try Periscope, which will use
sampling to speed up your analyses without any work on
your end!

Page 15

Thanks for Reading!

We hope you gained some new tools for your SQL toolbelt. We put a lot of work into our
first ebook and hope to put in even more as we expand on our techniques for making
SQL faster.
Please email us at hello@periscope.io with any questions or feedback about this ebook,
Periscope, or SQL in general. Wed love to hear from you!
Shameless Plug Alert
Looking for a better way to analyze your SQL database data? Want your queries to run
faster? Want to make sharing reports across your business easy?
Thousands of analysts use Periscope every day to dig deeper into their data and find the
insights they need to grow their business.
Customers realize an average of 150X improvement in query speeds with our
preemptive data cache. Even our world class support is fast, with an average response
time of 6 seconds during business hours.
Head to Periscope.io and sign up to get started with our 7 day free trial.
This is Just the Beginning
There are a number of techniques for optimizing queries that we didnt cover here. We
have a lot of ideas that didnt fit in the first version, but that were excited to get out over
the following months.
Want to be kept up to date on our ebook updates? Sign up at periscope.io/optimizing-sql
to get on the mailing list and hear about all the ways were improving the guide.

Page 16

Big Data and Hadoop-Sentiment Analysis Using Flume and Hive
No ratings yet
Big Data and Hadoop-Sentiment Analysis Using Flume and Hive
27 pages
The 10x Academy
No ratings yet
The 10x Academy
9 pages
Tips To Build An Outstanding Dev Portfolio
100% (1)
Tips To Build An Outstanding Dev Portfolio
15 pages
Simplified Way of Processing Large Data Using Chunk in Laravel
No ratings yet
Simplified Way of Processing Large Data Using Chunk in Laravel
4 pages
Context:: 1.need of Project 2.project Stages 3.project Requirements
No ratings yet
Context:: 1.need of Project 2.project Stages 3.project Requirements
7 pages
Tune Your Zabbix For Better Performance
No ratings yet
Tune Your Zabbix For Better Performance
31 pages
Datawarehousing
No ratings yet
Datawarehousing
71 pages
Reasons-To-Choose Php-Laravel-Framework
No ratings yet
Reasons-To-Choose Php-Laravel-Framework
16 pages
PROIDS: Probabilistic Data Structure Based Intrusion Detection System
No ratings yet
PROIDS: Probabilistic Data Structure Based Intrusion Detection System
13 pages
FOODIGO
No ratings yet
FOODIGO
11 pages
Adhd - Kahui Ako Waitemata Presentation
No ratings yet
Adhd - Kahui Ako Waitemata Presentation
30 pages
CODING DOJO Master Course Packet
No ratings yet
CODING DOJO Master Course Packet
24 pages
For More Details Visit: OR Whats App On +971 50 1878635
0% (1)
For More Details Visit: OR Whats App On +971 50 1878635
2 pages
La Ravel 4 Cookbook
No ratings yet
La Ravel 4 Cookbook
240 pages
React Router: Jogesh K. Muppala
No ratings yet
React Router: Jogesh K. Muppala
5 pages
Affiliate Marketing: by Alekya TVL DM-05-060
No ratings yet
Affiliate Marketing: by Alekya TVL DM-05-060
11 pages
GitHub Assignment Submission Guide
No ratings yet
GitHub Assignment Submission Guide
10 pages
Nagios Vs Zabbix
No ratings yet
Nagios Vs Zabbix
2 pages
Free PDF Ebook: Sending Mail With Attachment Using Laravel and Mailtrap - Io
No ratings yet
Free PDF Ebook: Sending Mail With Attachment Using Laravel and Mailtrap - Io
5 pages
Modul Framework Laravel PDF
No ratings yet
Modul Framework Laravel PDF
117 pages
Ebook Laravel 4
No ratings yet
Ebook Laravel 4
105 pages
2 Nagios
No ratings yet
2 Nagios
48 pages
Unit 5 - Data Science & Big Data - WWW - Rgpvnotes.in
No ratings yet
Unit 5 - Data Science & Big Data - WWW - Rgpvnotes.in
17 pages
Types of Email Marketing'
No ratings yet
Types of Email Marketing'
2 pages
Laravel Shop Tutorial #1 - 2
No ratings yet
Laravel Shop Tutorial #1 - 2
9 pages
Machine Learning Based Recommender System For E-Commerce
No ratings yet
Machine Learning Based Recommender System For E-Commerce
9 pages
Director Ecommerce Software Development in United States Resume Brendan Moran
No ratings yet
Director Ecommerce Software Development in United States Resume Brendan Moran
2 pages
Email Engine Guide
No ratings yet
Email Engine Guide
260 pages
The 5 Deadliest Mistakes That 90% of Affiliate Marketers Make
No ratings yet
The 5 Deadliest Mistakes That 90% of Affiliate Marketers Make
10 pages
Ajax Interview Questions and Answers Set
No ratings yet
Ajax Interview Questions and Answers Set
15 pages
Email Marketing Project
No ratings yet
Email Marketing Project
10 pages
Outline: How To Install Development Setup
No ratings yet
Outline: How To Install Development Setup
8 pages
Shopify's Big Data Platform
No ratings yet
Shopify's Big Data Platform
28 pages
Affliate Marketing Be Your Boss
No ratings yet
Affliate Marketing Be Your Boss
4 pages
Design Document Database
No ratings yet
Design Document Database
62 pages
SocialMediaLab Package Tutorial
No ratings yet
SocialMediaLab Package Tutorial
17 pages
Laravel 5 Latest Off Line Documentation
100% (1)
Laravel 5 Latest Off Line Documentation
407 pages
RC Res
No ratings yet
RC Res
5 pages
03 US Case Study Nestle Waters 20210616
No ratings yet
03 US Case Study Nestle Waters 20210616
7 pages
Graph Analytics PDF
No ratings yet
Graph Analytics PDF
13 pages
Digital Marketing Types
No ratings yet
Digital Marketing Types
11 pages
Iab-Affiliate-Marketing-Handbook 2016
No ratings yet
Iab-Affiliate-Marketing-Handbook 2016
29 pages
MongoDB Mongoosess
No ratings yet
MongoDB Mongoosess
31 pages
Cheat Sheet AngularJS
No ratings yet
Cheat Sheet AngularJS
5 pages
Javascript Exercises and Solutions
No ratings yet
Javascript Exercises and Solutions
7 pages
Complete Vue Mastery Course Guide
No ratings yet
Complete Vue Mastery Course Guide
11 pages
Wordpress Optimization Guide
No ratings yet
Wordpress Optimization Guide
77 pages
Frontend Development
No ratings yet
Frontend Development
19 pages
Zabbix Install Manual
No ratings yet
Zabbix Install Manual
6 pages
Data Visualization - Metabase Guide
0% (1)
Data Visualization - Metabase Guide
7 pages
Azure Cloud Intro
No ratings yet
Azure Cloud Intro
34 pages
Learning React Modern Patterns For Developing React Apps 2nd Edition Ebook PDF
No ratings yet
Learning React Modern Patterns For Developing React Apps 2nd Edition Ebook PDF
42 pages
C100dev 3
No ratings yet
C100dev 3
14 pages
N Tier Architecture
100% (1)
N Tier Architecture
31 pages
Thingworx Extension Development User Guide PDF
No ratings yet
Thingworx Extension Development User Guide PDF
93 pages
Android Ui Design
No ratings yet
Android Ui Design
58 pages
Optimizing Hadoop for MapReduce
From Everand
Optimizing Hadoop for MapReduce
Khaled Tannir
No ratings yet
Instant Redis Optimization How-to
From Everand
Instant Redis Optimization How-to
Arun Chinnachamy
No ratings yet
ColdFusion Interview Questions, Answers, and Explanations: ColdFusion Certification Review
From Everand
ColdFusion Interview Questions, Answers, and Explanations: ColdFusion Certification Review
equitypress
No ratings yet
Database testing Third Edition
From Everand
Database testing Third Edition
Gerardus Blokdyk
No ratings yet
Variables and Data Types in C#
No ratings yet
Variables and Data Types in C#
14 pages
Description of All Keywords in C
No ratings yet
Description of All Keywords in C
8 pages
Lab Manual Intro To ASP - Net, HTML
No ratings yet
Lab Manual Intro To ASP - Net, HTML
16 pages
Web Techonolog ETE QP
No ratings yet
Web Techonolog ETE QP
2 pages
Chap 2
No ratings yet
Chap 2
11 pages
AVL Tree Implementation in C
No ratings yet
AVL Tree Implementation in C
4 pages
Chapter2-Software Process
No ratings yet
Chapter2-Software Process
63 pages
Object-Oriented Database: Adoption of Object Databases
No ratings yet
Object-Oriented Database: Adoption of Object Databases
5 pages
Built in Data Type
No ratings yet
Built in Data Type
19 pages
File Organization and Indexing - DPP 01
No ratings yet
File Organization and Indexing - DPP 01
8 pages
Compiler Construction The Art of Niklaus Wirth
No ratings yet
Compiler Construction The Art of Niklaus Wirth
14 pages
MySQL Commands Cheat Sheet by PhoenixNAP
No ratings yet
MySQL Commands Cheat Sheet by PhoenixNAP
1 page
IS2103 A Dist Computing 1
No ratings yet
IS2103 A Dist Computing 1
18 pages
ADI Notes
No ratings yet
ADI Notes
5 pages
ETW-REST API v1.01 Scope and Sequence
No ratings yet
ETW-REST API v1.01 Scope and Sequence
2 pages
24msd7048(Python Lab-2).PDF (1)
No ratings yet
24msd7048(Python Lab-2).PDF (1)
12 pages
Pigz PDF
No ratings yet
Pigz PDF
3 pages
COMP2150 Winter2024 Test2 A01 Handout
No ratings yet
COMP2150 Winter2024 Test2 A01 Handout
2 pages
Eclipse IDE Macros
No ratings yet
Eclipse IDE Macros
17 pages
21MCA3033 - Worksheet 1.3.1
No ratings yet
21MCA3033 - Worksheet 1.3.1
5 pages
HP VuGen Certification
No ratings yet
HP VuGen Certification
4 pages
CIMATRON Tools
No ratings yet
CIMATRON Tools
18 pages
Development of A Rfid Based Door Access System 1
No ratings yet
Development of A Rfid Based Door Access System 1
71 pages
Python Project
No ratings yet
Python Project
12 pages
CSP Unit 3 App Development Planning Guidee
0% (1)
CSP Unit 3 App Development Planning Guidee
9 pages
Excellence Academy
No ratings yet
Excellence Academy
22 pages
Stream API
No ratings yet
Stream API
4 pages
Python Operators and Lists Completed Version
No ratings yet
Python Operators and Lists Completed Version
2 pages
Sample Python Programming
No ratings yet
Sample Python Programming
5 pages
1Z0-061 Sample Questions Answers PDF
No ratings yet
1Z0-061 Sample Questions Answers PDF
6 pages