turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

04-02-2012 10:27 PM

Hi,

I have a question in regard to finding the common elements in two vectors.

I have a column vector of tickers for a stock index. I have a loop over time which does some calculations at each day.

At each day inside the time loop, I have around 40 tickers, which are repeating again and again through time.

However, in some days new stocks may introduce to the index so the column vector of tickers at day t may have one of two tickers more than the column vector of tickers at day t-1.

For example, as it can be seen in the following table, at day t, NOP enters into the index for the first time. To observe these new stocks, I want to introduce a new variable as NEW which gives me zero for each common ticker and one for the new tickers which enter the index for the first time. So, how to define variable NEW in SAS/IML environment?

Your advice is appreciated.

DAY (T-1) | DAY (T) | NEW |
---|---|---|

AXA | AXA | 0 |

AIA | AIA | 0 |

BCD | BCD | 0 |

FEJ | FEJ | 0 |

JUD | JUD | 0 |

KID | KID | 0 |

LOD | LOD | 0 |

MIN | MIN | 0 |

OOP | NOP | 1 |

QRE | OOP | 0 |

TRE | QRE | 0 |

WTF | TRE | 0 |

XDE | WTF | 0 |

ZSW | XDE | 0 |

ZSW | 0 |

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

04-03-2012 03:24 AM

**One solution: **

data stocks;

input old $4. new $4.;

cards;

AXA AXA 0

AIA AIA 0

BCD BCD 0

FEJ FEJ 0

JUD JUD 0

KID KID 0

LOD LOD 0

MIN MIN 0

OOP NOP 1

QRE OOP 0

TRE QRE 0

WTF TRE 0

XDE WTF 0

ZSW XDE 0

ZSW 0

;

quit;

/* ~~~~~~~~~~~~~~~~~~~~~ IML Code ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ */

proc iml;

use stocks;

read all;

Valid_Rows = nrow( old[loc(old ^= '')] );

Valid_Old = old[loc(old ^= '')];

Mesh = repeat(new,1,valid_rows);

old_reshape = repeat(valid_old,1,nrow(mesh))`;

New_ID = (old_reshape = mesh)[,+] = 0;

print new New_ID;

quit;

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

04-03-2012 06:37 AM

I suggest using the ELEMENT function in SAS/IML 9.3.

To use Ajay's data:

/* b = elements of NEW that are elements of OLD */

b = element(new, old);

New_ID = ^b; /* change 0 <--> 1 */

print new New_ID;

If you don't have SAS 9.3, I can suggest another way.

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

04-03-2012 06:57 AM

Thanks,

Yes please, I am using SAS 9.2.

For further clarification, I have column vectors DAY(T) and DAY(T-1) and want to produce the column NEW by having a loop over DAYs columns.

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

04-03-2012 08:10 AM

Here's a SAS/IML module that duplicates the functionality of the ELEMENT function that was introduced in SAS/IML 9.3:

start elementMod(x, y);

if ncol(x)=0 then return(_NULL_);

b = j(nrow(x), ncol(x), 0);

if ncol(y)=0 then return(b);

u = xsect(x, y);

if ncol(u)=0 then return(b);

do i = 1 to ncol(u);

b = bor(b, (x=u*));*

end;

return (b);

finish;

New_ID2 = ^elementMod(new, old);